Investigating Psychosocial Factors: Supporting Clinical Decisions for Outpatient Diabetes Care

by

Charles R. Senteio

A dissertation submitted in partial fulfillment of the requirements for the degree of

Doctor of Philosophy (Information)

in the University of Michigan 2015

Doctoral Committee:

Associate Professor Tiffany Veinot, Chair Assistant Professor Julia Adler-Milstein Associate Professor Caroline R. Richardson Associate Professor Kai Zheng

© Charles R. Senteio 2015

ii

To Mother and Dad Happy are those who dream dreams, and are willing to pay the price to make them come true.

In pursuit of our higher ground, indeed, we find our path by walking it.

iii

ACKNOWLEDGEMENTS

This dissertation has achieved one measure of success by answering my research

questions, which were refined and honed through the tireless support of the dissertation

committee, several other researchers, colleagues, and other health care practitioners. It is

also successful in that it helps fulfill a personal goal, conceived over a decade ago. At that

time, I began to understand the chasm that existed between practitioners and certain at-

risk patients. Over the course of hundreds of home visits and clinical consultations, I

witnessed at-risk patients lament that the health care “system” failed to understand their

plight; I also had the opportunity to work with highly skilled practitioners who expressed

frustration with not knowing enough about—or learning too late of—the persistent

structural and individual circumstances that afflicted certain patients who simply could

not overcome various barriers to care. At that time, I committed to a journey to do what I

could to help close that chasm, to participate meaningfully in extending the capabilities of

“personalized” medicine, and to better incorporate the patient’s social and environmental

experience into the chronic care clinical consultation to facilitate better outcomes. This

dissertation is an important step in that journey; it represents the last, major academic

milestone in my time as a doctoral student at the University of Michigan School of

Information, time that has included earning a Master’s degree from the School of Social

Work. I am extremely fortunate to have had the support of many individuals that made

this step possible.

iv

I deeply appreciate the support and guidance from my dissertation chair and

advisor, Dr. Tiffany Veinot. I still recall our conversation in April, 2011 when I told you I

would be coming back to Michigan for my doctoral studies. I would be returning to the

university that had already played a significant role in my personal and professional

development as a Master’s student in Business Administration, then as an alumnus. I

appreciated the joy you conveyed through the phone when I informed you of my

decision. I noted that same enthusiasm when you informed me of the honor of being

awarded the Gary M. Olson Outstanding Ph.D. Student Award. You have been an

influential presence, a true scholar, who has been consistent in your tireless efforts to

support my development as a researcher. I appreciate you helping me to expand my

network, through introductions to the wonderful people at the Veterans Administration,

faculty at other institutions, and members of your network dedicated to executing quality

research work and translating it into practice. Thank you especially for connecting me to

special people like Terrance Campbell, Dave Law, and Tammy Toscos.

I established three simple criteria when I began contemplating potential members

of my dissertation committee: 1) proven scholars who could help augment my research

needs, 2) ability to work well with me, and each other, and 3) clear demonstration of

genuine passion for their work. Each of you have gone well beyond my expectations for

providing me the level and type of support I sought. Thank you. Julia Adler-Milstein, I

had the privilege to meet and get to know you when I took your health policy course in

the fall of 2011, my first semester at the School of Information. You have been a source

of constant support and pragmatic counsel, no matter the type of question or issue I

brought to you. Thank you, Kai Zheng, for consistently responding to my requests for

v

support. You, like Tiffany and Julia, were part of my comprehensive exam, and

dissertation committee. I will always appreciate your support for my dissertation and my

other work in healthcare. Caroline Richardson, you have been a constant source of

support since I first asked if you would be willing to serve on my dissertation committee.

Your clinical perspectives have been invaluable in helping to shape this study. You also

have granted me access to your network of other physicians. That access has greatly

enhanced this work. I will always cherish the time and effort each of you provided,

essential in enabling me to complete this work.

There are several other researchers who have helped me execute this work. Jim

Walton, your skill as a clinician and strategic thinker helped me shape this work long

before I started the doctoral program; access to your organization was indispensable to

this effort. Thank you for your friendship. Julie Lowery. I will always cherish our spirited

debates about the relevance of this work. You have pushed me even further in helping me

shape my own priorities for translational research. Your spirit and competence continue

to inspire. Thank you, James Dickens, Badia Harlin, and Martha Funnell for offering

your rich insights on non-physician practitioners and diabetes care. Furthermore, I deeply

appreciate your considerable efforts in providing me access to the professional

organizations you lead.

I would also like to express my gratitude for individuals at the Veterans

Administration in Ann Arbor who provided support and guidance throughout this project:

Tammy Chang, Alicia Cohen, Elissa Gaies, and Beverly Hall. In addition, my

involvement in the Disparities Leadership Program at Massachusetts General Hospital

provided me with the opportunity to work with, and learn from, national leaders who

vi

work with at-risk patient populations. A very special thanks to: Aswita Tan-McGrory,

Joseph R. Betancourt, Liz Trevino, Derrick Villa, Cunegundo “Connie” Vergara, Zoila

Torres Feldman, Gavin Muir, and Kris McCracken.

I am very pleased to acknowledge the faculty at the School of Information who

have been a tremendous source of support and guidance: Paul Edwards, Mark Ackerman,

David Wallace, John King, Lionel Robert, Olivia Frost, Jeff MacKie-Mason, Tom

Finholt, Eric Cook, Kristin Fontichiaro, Chuck Friedman, Tawanna Dillahunt, Stephanie

Teasley, and Beth Yakel. I also would like to express my gratitude to School of

Information staff for their support over the past four years: Beth Zibro, Judy Lawson,

Barb Smith, Veronica Falandino, Jay Jackson, Ben Armes, Glenda Bullock, Sue Schoen,

Mike Doa, Lai Tutt, Jim Schmidt, Heather Carpenter, Heather Newman, Rebecca

O’Brien, Sheryl Smith, Lorraine Robert, Shamille Orr, and Kathleen O’Connor.

I have been privileged to be a part of a student community that is as supportive as

any I have known. I want to extend a very special thanks to the School of Information

students who have been so supportive: Devan Donaldson, Pablo Quinones, Andrea

Barbarin, Melody Ku, Tawfiq Ammari, Karina Kervin, Chris Wolf, Danny Wu, Allen

Flynn, Christa Meadowbrooke, Elizabeth Kaziunas, Adam Kriesberg, Gaurav Paruthi,

Carrie Xu, Melissa Chalmers, Rayoung Yang, Xin Rong, Jasmine Jones, and a very

special thank you to Ms. Lois Street.

There are other University of Michigan faculty, staff and students I would like to

thank for their support in helping to make this experience so fulfilling. I will cherish your

guidance and support. Thank you, Diane Vinokur and Jerry Davis. You both have served

as wonderful, familiar sounding boards for so many of my ideas over these many years.

vii

You helped me transition back to Ann Arbor as a full time student. Other University of

Michigan faculty and staff have provided me with advice and support as I worked

through the program: Greg Merritt, Renee R. Anspach, Tim McDaniel, Daphna

Oyserman, Daniel Keating, Sofia Merajver, Bob Anderson, Bob Quinn, and Shyamala

Nagaraj. Also, faculty and staff at the School of Social Work offered me endless support:

Mike Spencer, and Ricardo Guzman, Laura Sanders, Karen Staller, William Vanderwill,

Sue Crab, and Leslie Hollingsworth. Thank you to those I had the tremendous privilege

to meet and get to know through the Dean of Students Advisory Board: Dean Laura

Jones, Rachel Naasko, Andrew Kruger, Luciana Aenășoaie, and Bro. Canon Thomas.

I would also like to express my deepest thanks to two individuals I had the

privilege to meet and get to know since starting the doctoral program: David Blackwolf

Belcourt (1970 – 2014) and Noreen Clark (1943 – 2013). Your kindness and passion for

serving others endures and continues to inspire me.

My research has been made possible by the generous support from the School of

Information and the Rackham Graduate School. The Rackham Merit Fellowship award

enabled me to focus on my research and gave me the freedom to pursue opportunities to

develop as a researcher and scholar. A special thanks to the Rackham faculty and staff for

their support: Mark Kamimura-Jimenez, Bro. Larry Rowely, Emma Flores, Natalie

Bartolacci, Arahshiel Silver, and Regina Sims.

I feel extremely privileged to have had the opportunity to experience the

University of Michigan student community. One of my goals in returning to Ann Arbor

full time was to try and get to know students outside my program. These individuals

welcomed me into a very special community of scholars across the campus: Babe

viii

Kawaii-Bogue, Elise Hernandez, Joseph Bayer, Jorge Solar, William Lopez, Jeremy

Lapham, Lillian Hanan Al-Bilali, Yet Renée, Peter Jacobs, Rebecca Cheezum, Charity

Hoffman, Austin Dixon, Shannon Stone, Mohammad Iskandarani, Katy Downs, Ikenna

Nwamba, Armando Bernal, Scott Kalafatis, Winstone Seow Wei Shen, and Wes Darland.

I want to thank the very special people in Texas who provided endless support as

I embarked on this journey north. Thank you for helping me hold things down: Bro.

Victor Edwards, Stacy Welk, Pastor James Reed, Pastor Preston Weaver, Bro. Ray

Schufford, Phyllis Jackson, Nash Haggerty, Kartieaa Coleman, Mark Edwards, Marisela

Rosales, Bro. Dale Long, Katie Cardarelli, Marcus Martin, Rev. Donald Parish, Karen

Pettis, Daniel Perez, and Justin Waninger.

I want to acknowledge my very special, longstanding friendships that provided

the moral support I needed to do this. Thank you, Bro. Alfy Roby, Bro. Dwayne Ellis,

Sean Hudson, Bro. DeLance Farrell, Vincent Harris, Mark Brown, Sydney Levy,

Kathryne Smith, Greg Stewart, and Robert LaBril. Thank you also to the folks whose

talent and passion helped me get through my early morning rituals: Mos Def, Rafael

Saadiq and “Krsna” Parker.

I will conclude with a very special thank you to my family who has always

supported me in pursuit of my own significance. Thank you, mother and pops. You have

given me the wonderful freedom to grow and explore, with a structure in which to do so.

Indeed, here on Earth, God’s work must truly be our own. I have only known love and

support from you, and that is a big part of what motivates me. Charlene, thank you for

your support, especially during this program. I appreciate you helping me stay focused on

what is really important.

ix

This step represents a part of a journey that began before I entered the doctoral

program, which I hope will persist. These very special individuals will always be a part

of this important step. I continue to believe that enhancing capabilities to collect and use

psychosocial information can achieve two important objectives: to 1) arm passionate,

skilled clinicians with the information they need to provide more effective care, and 2)

concurrently provide more effective channels for patients to share their lived experience.

This dissertation represents an important step toward that end, and I look forward to

continuing the journey!

x

Table of Contents DEDICATION................................................................................................................... ii ACKNOWLEDGEMENTS ............................................................................................ iii LIST OF TABLES ......................................................................................................... xiii LIST OF FIGURES ........................................................................................................ xv

LIST OF APPENDICES ............................................................................................... xvi ABSTRACT ................................................................................................................... xvii CHAPTER 1 INTRODUCTION ..................................................................................... 1 1.1. Overview ............................................................................................................................ 2 1.2. Research Questions ............................................................................................................ 2 1.3. Study Design ...................................................................................................................... 3 1.4. Results and Contribution .................................................................................................... 3

CHAPTER 2 LITERATURE REVIEW ......................................................................... 8 2.1. Diabetes Context ................................................................................................................ 8 2.2. Literature Review Overview – Conceptual Framework ..................................................... 9 2.3. Psychosocial Factors - Overview and Influence on T2DM Self-Care ............................. 14 2.3.1. Psychosocial Factors and Health Outcomes ................................................................ 16 2.4. Clinical Practice Guidelines (CPGs) in Support of Clinical Decisions ............................ 25 2.4.1. T2DM Outcomes Evaluated Against CPGs ................................................................. 28 2.5. Clinical Decision Making ................................................................................................. 31 2.5.1. Clinical Information Access for Clinical Decision Making ......................................... 36 2.6. Gaps in Understanding Psychosocial Information Access and Use ................................. 39

CHAPTER 3 RESEARCH DESIGN ............................................................................ 46 3.1. Research Plan / Approach ................................................................................................ 46 3.2. Timeline ........................................................................................................................... 47 3.3. Study Population, Sampling Plan & Participant Recruiting Methods .............................. 48 3.4. Data Collection ................................................................................................................. 53 3.4.1. In-Depth, Semi-Structured Interviews ......................................................................... 53 3.4.2. Survey Design .............................................................................................................. 54 3.4.3. Survey Distribution and Administration ...................................................................... 57 3.5. Data Analysis ................................................................................................................... 58 3.5.1. Interview Analysis ....................................................................................................... 59 3.5.2. Survey Data Analysis ................................................................................................... 61 3.6. Risk Management ............................................................................................................. 64

CHAPTER 4 ACCESSING PSYCHOSOCIAL INFORMATION FOR CLINICAL DECISIONS..................................................................................................................... 66 4.1. Introduction ...................................................................................................................... 66 4.2. Characteristics of Study Participants ................................................................................ 67 4.3. Overview of Findings ....................................................................................................... 70

xi

4.4. Importance of Psychosocial Factors ................................................................................. 73 4.4.1. Physician Interview Findings ....................................................................................... 74 4.4.2. Survey Findings ........................................................................................................... 84 4.4.3. Comparison of Results ................................................................................................. 86 4.5. Sources of Psychosocial Information ............................................................................... 87 4.5.1. Physician Interview Findings ....................................................................................... 88 4.5.2. Survey Findings ........................................................................................................... 92 4.6. Practitioner Characteristics – Physician Interview Findings ............................................ 93 4.6.1. Survey Results ............................................................................................................. 95 4.7. Conclusion ........................................................................................................................ 96

CHAPTER 5 USING PSYCHOSOCIAL INFORMATION FOR CLINICAL DECISIONS..................................................................................................................... 98 5.1. Introduction ...................................................................................................................... 98 5.2. Cognitive Map of Psychosocial Information Use .......................................................... 100 5.3. Considering Clinical Practice Guidelines (CPGs) in Context of Patient Situation ........ 101 5.4. Build & Maintain Rapport with Patient.......................................................................... 105 5.4.1. Techniques to Build and Maintain the Patient Relationship ...................................... 109 5.5. Triggers to Gathering and Using Psychosocial Information .......................................... 111 5.6. Patient Assessment ......................................................................................................... 124 5.7. Making the Clinical Decision ......................................................................................... 128 5.7.1. Medication Management ............................................................................................ 128 5.7.2. Recommendations ...................................................................................................... 136 5.7.3. Determining Target Level of Control ........................................................................ 140 5.7.4. Referrals ..................................................................................................................... 144 5.7.5. Survey Findings – Influence on Clinical Decisions ................................................... 148 5.8. Conclusion ...................................................................................................................... 154

CHAPTER 6 BARRIERS AND FACILITATORS TO USE OF PSYCHOSOCIAL INFORMATION ........................................................................................................... 156 6.1. Introduction .................................................................................................................... 156 6.2. Facilitators to Use ........................................................................................................... 157 6.2.1. EHR Can Facilitate Use ............................................................................................. 157 6.2.2. Survey Results ........................................................................................................... 159 6.3. Barriers to Use ................................................................................................................ 160 6.3.1. Clinical Practice Constraints ...................................................................................... 160 6.3.2. Concerns about Consistency Across the Care Team .................................................. 163 6.3.3. Questions Regarding Accuracy of Psychosocial Information Accessed.................... 164 6.3.4. EHR Presents Barriers to Use .................................................................................... 167 6.3.5. Availability of Psychosocial Information to Support Clinical Decisions .................. 171 6.4. Final Conceptual Model of Psychosocial Information Access ....................................... 173 6.5. Additional Psychosocial Information Desired ................................................................ 174 6.6. Improvement Ideas ......................................................................................................... 177 6.7. Survey Findings – EHR Tools ........................................................................................ 178 6.7.1. Frequency with which EHR Tools Support Documentation and Retrieval ............... 179 6.7.2. Confidence in Using EHR Tools ............................................................................... 180 6.8. Conclusion ...................................................................................................................... 181

CHAPTER 7 DISCUSSION AND IMPLICATIONS ................................................ 185 7.1. SUMMARY OF FINDINGS .......................................................................................... 185 7.2. Introduction .................................................................................................................... 191 7.3. Four Major Findings ....................................................................................................... 193

xii

7.4. Strengths and Limitations ............................................................................................... 206 7.5. Conclusion ...................................................................................................................... 209

APPENDICES ............................................................................................................... 211

REFERENCES .............................................................................................................. 257

xiii

LIST OF TABLES

Table 2.3 Psychosocial Factors that Affect T2DM Outcomes...................................15

Table 3.4.2 Cognitive Interview Participants................................................................57

Table 3.4.3 Survey Distributions ..................................................................................58

Table 3.5.2 Survey Response Rates ..............................................................................63

Table 4.2.1 Interview Participant Characteristics .........................................................68

Table 4.2.2 Professional Characteristics of Survey Respondents .................................69

Table 4.2.3 Demographic Characteristics of Survey Respondents ...............................70

Table 4.4.2 Average Likert Scores of Individual Psychosocial Factors .......................86

Table 4.5.1 Sources of Psychosocial Information, How Accessed ...............................87

Table 4.5.2 Sources of Psychosocial Information.........................................................92

Table 5.5 Triggers Gathering and Using Psychosocial Information ........................113

Table 5.5.1 Chronic Circumstance ..............................................................................114

Table 5.5.2 New Circumstance ...................................................................................118

Table 5.5.3 Change in Circumstance ..........................................................................121

Table 5.5.4 Circumstances when Psychosocial Factors are Important to Consider ...123

Table 5.6 Psychosocial Information Use for Patient Assessment ............................124

Table 5.7.1 Psychosocial Information Use for Medication Management...................129

Table 5.7.2 Psychosocial Information Use for Recommendations .............................137

Table 5.7.3 Psychosocial Information Use for Target Levels of Control ...................141

Table 5.7.4 Psychosocial Information Use for Referral Decisions .............................144

Table 5.7.5.1 Frequency with which Decisions are Influenced by Psychosocial Factors

..................................................................................................................148

Table 5.7.5.2 Frequency with which Level of Control Decisions are Influenced by

Psychosocial Factors ................................................................................149

Table 5.7.5.3 Frequency with which Recommendations Decisions are Influenced by

Psychosocial Factors ................................................................................150

xiv

Table 5.7.5.4 Frequency with which Referrals Decisions are Influenced by Psychosocial

Factors ......................................................................................................151

Table 5.7.5.5 Frequency with which Medications Decisions are Influenced by

Psychosocial Factors ................................................................................152

Table 6.2.2 Confidence in Accuracy of Psychosocial Information ............................160

Table 6.3.5 Frequency with which Respondents Have the Psychosocial Information

They Need ................................................................................................172

Table 6.7.1.1 Frequency that EHR Tools Support Documentation of Psychosocial

Information ..............................................................................................179

Table 6.7.1.2 Frequency that EHR Tools Support Retrieval of Psychosocial Information

..................................................................................................................180

Table 6.7.2.1 Confidence in Using EHR Tools to Support Documentation of

Psychosocial Information.........................................................................180

Table 6.7.2.2 Confidence in Using EHR Tools to Support Retrieval of Psychosocial

Information ..............................................................................................181

xv

LIST OF FIGURES

Figure 2.2: Initial Conceptual Framework of Psychosocial Factors in Outpatient

Diabetes Care ................................................................................................9

Figure 2.2.1: IOM Table of Psychosocial Factors ............................................................11

Figure 2.6: Conceptual Framework of Psychosocial Factors in Outpatient Diabetes

Care .............................................................................................................44

Figure 3.1: Exploratory Sequential Study Design .........................................................47

Figure 3.2: Study Timeline ............................................................................................48

Figure 3.4.2: Survey Design Process ................................................................................55

Figure 4.3: Initial Conceptual Model of Psychosocial Information Access ..................72

Figure 5.2: Cognitive Map of Psychosocial Information Use .....................................101

Figure 6.4: Final Conceptual Model of Psychosocial Information Access .................174

Figure 7.2: Final Conceptual Framework of Psychosocial Factors in Outpatient

Diabetes Care ............................................................................................192

xvi

LIST OF APPENDICES

Appendix A: Definitions ..................................................................................................212

Appendix B: Psychosocial Factors ..................................................................................213

Appendix C: Type 2 Diabetes Clinical Decisions ...........................................................215

Appendix D: Outline of ADA Standards of Care ............................................................216

Appendix E: Interview Participants .................................................................................218

Appendix F: Emails Used for Survey Distribution ..........................................................219

Appendix G: Semi-Structured Interview Guide ..............................................................231

Appendix H: Survey Instrument ......................................................................................233

Appendix I: Code Book for Physician Interviews ...........................................................250

Appendix J: Principles of Survey Instrument Design ......................................................254

Appendix K: Average Likert Scores of Psychosocial Factors .........................................255

Appendix L: Average Likert Scores of Psychosocial Factors by Group .........................256

xvii

ABSTRACT

Practitioners often make decisions to diverge from clinical targets specified in

diabetes clinical practice guidelines (CPGs) based on each patient’s unique situation,

although reasons for this are poorly understood. Additionally, while practitioners

understand that psychosocial factors may influence self-care behavior and subsequently

health outcomes, little is known as to how psychosocial factors influence clinical

decisions, including those that deviate from guidelines. Therefore, through a sequential

exploratory mixed methods study involving physician interviews (n=17) and a survey of

physicians, nurses and diabetes educators (n=229) I investigated practitioners’ access to,

and use of, psychosocial information as a basis for clinical decisions in outpatient

diabetes care. This study revealed four major findings. First, that psychosocial

information is not considered when patients have good glycemic control, but they do

consider it when a patient: 1) has persistent, poor glycemic control, 2) is a new patient or

has a new diabetes diagnosis, and 3) worsening of glycemic control. Second, access to

psychosocial information is granted through dialogue in an ongoing, trusting relationship.

Physicians use specific techniques to build trusting relationships, which include

demonstrating caring and creating a safe environment characterized by patient autonomy

and privacy. Third, awareness of psychosocial information may trigger decisions to

personalize HbA1c targets, pursue less aggressive treatment plans or augment guideline-

concordant treatment with actions to address barriers to care, such as referrals to

xviii

prescription assistance. Fourth, EHR designs are not optimized for capturing and

retrieving qualitative and situationally-dependent psychosocial information, which tends

to come in a narrative form. Specifically, study findings offer new insight into

circumstances in which practitioners’ decisions may deviate from CPGs, and their

rationales for doing so. Practitioners connecting patients to supplemental resources

represent efforts to reduce negative impacts of psychosocial factors on diabetes-related

self-care. If successful, these actions could ultimately improve diabetes outcomes.

Findings regarding the importance of a trusting clinician-patient relationship also suggest

the importance of care continuity to psychosocial information use. Moreover, results

indicate that effective use of psychosocial information requires unique socio-technical

supports that include clinician-patient relationship-building efforts and digital tools that

are optimized for the capture and retrieval of information in narrative form.

1

CHAPTER 1

INTRODUCTION

The dominant model of disease today is biomedical, and it leaves no room within its framework for the social, psychological, and behavioral dimensions of illness.

― George L. Engel, 1977

Psychosocial factors influence outcomes for adult, type 2 diabetes mellitus

(T2DM or “diabetes”) patients; such factors affect patient self-care practices which are a

vital component of the diabetes treatment regimen. “Psychosocial factors” is a general

term used in various areas of healthcare research. In this study, I define psychosocial

factors as the psychological factors—how an individual thinks and feels—and social

factors—an individual’s social milieu—that affect self-care behavior. They are the

individual (e.g., financial circumstance, perceptions) and structural (e.g., social support,

community resources, and cultural traditions) factors that influence self-care behavior

(see Appendix B: Psychosocial Factors) (Brotman, Golden, & Wittstein, 2007; Bruner et

al., 2004; Institute of Medicine, 2008a; Kemp & Brandwein, 2010; Macleod & Davey

Smith, 2003; Martikainen, Bartley, & Lahelma, 2002; McEwen, 1998; Singh-Manoux,

2003). Psychosocial information differs from clinical information which is currently

regularly collected and used (i.e., information contained in lab reports) in two ways: 1) it

tends to be qualitative in nature, and 2) it is situational, defined by the patient’s

circumstances. Although psychosocial factors are widely recognized to affect patient

adherence to recommended diabetes self-care behaviors, little is known about how health

2

care practitioners, such as physicians and nurses, consider such factors as they make, or

provide input into, various diabetes clinical decisions (see Appendix C :Type 2 Diabetes

Clinical Decisions). Further, information use models do not explain how psychosocial

factors may influence decisions to deviate from practice guidelines, which are important

considerations when evaluating practice performance against clinical goals.

1.1. Overview This study investigates how health care practitioners access and use psychosocial

information to provide diabetes care. Specifically, I examine the relevance of

psychosocial factors, the sources of psychosocial information, and the particular clinical

decisions that are influenced by taking these factors into consideration. Further, I describe

elements that impede and facilitate the use of psychosocial information, including the

role that current electronic health record (EHR) tools play in the documentation and

retrieval of psychosocial information.

1.2. Research Questions This study has been organized to answer the following research questions:

1. Which psychosocial factors do practitioners perceive as important in making, or

providing input into, care decisions for adult, type 2 diabetes patients? What is

their relative priority?

2. How do practitioners access psychosocial information?

3. How do practitioners use psychosocial information? How does this information

influence their care decisions?

4. In which situations are psychosocial factors considered?

5. What practitioner characteristics (i.e. role, age) are associated with their use of

psychosocial information?

3

6. What are the barriers and facilitators to acquiring and using psychosocial

information? How effectively do current tools (templates, data fields, free text)

support the storage and retrieval of psychosocial information?

1.3. Study Design I used a mixed methods design in order to leverage the strengths of both

qualitative and quantitative approaches. Initially, I conducted seventeen in-depth, semi-

structured interviews with physicians with experience providing diabetes care in the

outpatient setting, primarily at sites where the physicians practiced. I then developed an

online survey instrument, informed by my analysis of the interview data. To determine

generalizability and explore relationships between variables, I then administered the

online survey to primary care physicians, nurse practitioners, and diabetes educators.

Survey participants are individuals who, depending on their specific clinical role, have

experience making, or providing input into, diabetes care clinical decisions in the

outpatient setting.

I used a grounded theory approach to analyze the interview data. This analysis

included developing categories and constructing visuals of findings that served as the

foundation of a cognitive map, which depicts the physicians’ thinking as they use

psychosocial information in the course of providing diabetes care. I used descriptive

statistics to describe the size and distributions of the various elements of the survey

sample and inferential statistics to measure associations between variables, specifically

independent t tests of differences in means, and a logistic regression.

1.4. Results and Contribution There are four major findings resulting from this investigation, gleaned from both

the interview and survey data analysis.

4

First, analysis of the interview and survey data revealed that practitioners do not

consider psychosocial information when patients are stable and well controlled, but it is

considered under three specific circumstances: 1) under chronic circumstances such as

when treating at-risk patients or when a patient experiences persistent, poor glycemic

control; 2) under new circumstances, such as when seeing a new patient or a patient who

has been newly diagnosed with diabetes; and 3) when there is a change in circumstances,

such as when a patient experiences worsening of glycemic control and/or when there is a

sudden increase in unhealthy self-care behavior.

Second, analysis of the interview data revealed that physician’s access to

psychosocial information is granted through dialogue in the context of an ongoing

trusting relationship. Patients are the most frequent source of psychosocial information,

and study participants indicate that patients grant access to this information only if the

relationship has been established and is maintained. They establish trust by involving the

patient in clinical care decisions and respecting the patient’s privacy. They also avoid

pejorative language and tone.

Third, analysis of the interview data revealed that awareness of psychosocial

information may trigger decisions to personalize HbA1c targets, pursue less aggressive

treatment plans or augment guideline-concordant treatment with actions to address

barriers to care. For example, these actions include referrals to: prescription assistance,

food support, counseling, and transportation aid. Such actions can be triggered by

awareness of a patient experiencing: financial strain, mental health issues, low social

support, or other issues.

5

Fourth, analysis of the interview and survey data revealed that current electronic

health record (EHR) designs are not optimized for capturing and retrieving qualitative

and situationally dependent psychosocial information which tends to come in narrative

form. In addition, survey participants indicated that, among the sources I investigated, the

EHR is the least frequently used and viewed to be the least reliable as illustrated by the

participants’ low confidence in its accuracy. Notably, survey findings also indicate that

clinical decisions are occasionally made without the necessary psychosocial information.

The results of the study include a description of how psychosocial information is

considered in the process of making, or providing input into, diabetes care clinical

decisions. As part of this analysis, I created the conceptual model of psychosocial

information access (Figure 6.4) to describe how psychosocial information is accessed.

The model is based on both survey and interview data, and it includes the psychosocial

factors that practitioners perceive to be the most relevant, practitioners’ sources for this

information, the perceived accuracy and availability of psychosocial information, and the

associated clinical decisions that this information influences. The most relevant of the

twenty-three psychosocial factors examined are: financial strain, mental health status, life

stressors, food security, social support, and health literacy. This psychosocial information

is obtained primarily from the patient, the family and caregivers, other practitioners. The

EHR is also a source for this information, albeit a less important one.

I created the aforementioned cognitive map of psychosocial information use

(Figure 5.2) to describe how psychosocial information is used to make diabetes care

clinical decisions. The map represents how physicians from my interview sample use

psychosocial information to inform clinical decisions. It contains the following five

6

components: 1) Consider Clinical Practice Guidelines (CPGs) in Context of Patient

Situation, 2) Build and Maintain Rapport with Patient, 3) Making the Clinical Decision,

4) Assessing the Patient, and 5) Triggers to Gathering and Using Psychosocial

Information. The map depicts how psychosocial information informs clinical decisions;

this information helps physicians determine the care regimen appropriate for the patient

which is dependent partly upon relevant psychosocial factors.

Study results can be used to help guide future work concerning the use of

psychosocial information in three key research areas that investigate approaches to

improve chronic care outcomes: 1) influence of psychosocial factors on chronic disease

self-care behavior, 2) clinical decision making, and 3) use of clinical practice guidelines

(CPGs). I introduce a conceptual framework in chapter two (Figure 2.2) to situate the

focus of the investigation in the current literature, showing associations to these three

areas of research. I revisit and finalize the conceptual framework in chapter seven (Figure

7.2), where I discuss my major findings within the context of gaps in the literature. My

findings help advance the understanding of when and how psychosocial information is

used for diabetes care in the outpatient setting. They offer key insights regarding the

rationale physicians use to intentionally diverge from targets established in the CPGs.

Psychosocial information helps them understand that following the guidelines may not be

appropriate, and in some cases, may put the patient at risk. Findings also offer insights as

to how practitioners make clinical decisions to mediate the influence of psychosocial

factors when particular factors may present barriers to following recommended self-care

practices. Last, study findings offer key insights as to the practitioner understanding of

the influence of psychosocial factors on diabetes self-care. They include perspectives

7

based on three practitioner roles: physician, nurse practitioner, and diabetes educator. All

are responsible for making, or providing input into, diabetes care decisions.

In summary, study findings can be used by practitioners and developers of health

informatics capabilities (i.e. clinical decision support systems, electronic health records)

focused on improving diabetes outcomes. They can help support important

recommendations to build upon current capabilities to more effectively capture and

enable use of psychosocial information, which is often in narrative form. The

recommendations are focused on improving diabetes outcomes, recommendations for

which there is considerable, building momentum signified by the development of the

patient-centered medical home model (PCMH) and the Institute of Medicine detailing the

need to enhance the collection and use of psychosocial information, which I detail in the

next chapter.

8

CHAPTER 2

LITERATURE REVIEW

Millions of our citizens do not now have a full measure of opportunity to achieve and to enjoy good health. Millions do not now have protection or security against the economic

effects of sickness. And the time has now arrived for action to help them attain that opportunity and to help them get that protection.

― Harry S. Truman, September 19, 1945

2.1. Diabetes Context Type 2 diabetes mellitus (T2DM or “diabetes”) is a metabolic disorder of elevated

glucose concentration in blood. This chronic condition is caused by the gradual,

deleterious decline in the efficacy of the insulin hormone. As of the end of 2011—the

most current year the Centers for Disease Control and Prevention (CDC) has reported

statistics—there were nearly 21 million people with diabetes in the United States,

approximately 7.8 percent of the population. Since 1980, the number of Americans

diagnosed with T2DM has more than tripled (Center for Disease Control, 2011). Every

year, 1.3 million individuals are diagnosed. Escalating incidence and prevalence are

associated with self-care behavior. The number of new cases is associated with increased

prevalence of obesity; 80 percent of individuals with diabetes are either overweight or

obese (U.S. Department of Health and Human Services, 2013). The growing prevalence

of obesity is related to unhealthy dietary practices and sedentary lifestyle (Rodbard et al.,

2009).

9

2.2. Literature Review Overview – Conceptual Framework This literature review is organized to reflect the published literature in three areas:

psychosocial factors and diabetes self-care, clinical decision making (CDM), and clinical

practice guidelines (CPGs). I constructed a conceptual framework to situate the focus of

my research—practitioner perceptions and use of psychosocial information for diabetes

care—in the literature to illuminate the gaps that this research fills (see Figure 2.2)

(Ravitch & Riggan, 2011). The black box on the Initial Conceptual Framework of

Psychosocial Factors in Outpatient Diabetes Care is the central focus of this work,

showing relationships to the extant literature. In the first part of the literature review

(section 2.3), I describe the literature on the influence of psychosocial factors on self-care

which impacts diabetes outcomes. Second, I outline the literature on clinical practice

guidelines (CPGs) in support of clinical decisions (section 2.4). Third, I describe the

literature on clinical decision making, focusing on patient-centered care as this is the area

most aligned with use of psychosocial information specific to the patient situation

(section 2.5). As is evident from this Figure, there are currently gaps in understanding

practitioner perceptions and use of psychosocial information in providing diabetes care in

the outpatient setting. It is these gaps that the present research aims to fill.

* - PFs – Psychosocial Factors; * - CPGs – Clinical Practice Guidelines

Figure 2.2: Initial Conceptual Framework of Psychosocial Factors in Outpatient Diabetes Care

10

Psychosocial Factors Psychosocial factors are important because they influence diabetes outcomes,

primarily through their effect upon recommended self-care behavior. The diabetes care

regimen requires substantial responsibility from the patient for regular self-care practices;

adhering to them results in positive health outcomes. Psychosocial factors can both

facilitate, and present barriers to, diabetes self-care behavior.

For example, high levels of social support are consistent with specific healthy self-

care practices—dietary behavior, foot care, fasting blood sugar (FBS) testing and daily

physical activity (Tang, Brown, Funnell, & Anderson, 2008; Watkins, Quinn, Ruggiero,

Quinn, & Choi, 2013)—and ultimately, positive outcomes such as glycemic control (Iida,

Parris Stephens, Rook, Franks, & Salem, 2010; Mayberry & Osborn, 2012; Rintala,

Jaatinen, Paavilainen, & Åstedt-Kurki, 2013; Sharfi, Azad, Avat, & Siamak, 2013; Shier,

Ginsburg, Howell, Volland, & Golden, 2013). Also, spirituality has been shown to

positively influence social support (Polzer & Miles-Shandor, 2005, 2007; Quinn, Cook,

Nash, & Chin, 2001), and subsequently glycemic control (Newlin, Melkus, Tappen,

Chyun, & Koenig, 2008).

Conversely, financial strain has been shown to present barriers to medication

adherence, which lead to poor glycemic control (Dubois, Chawla, Neslusan, Smith, &

Wade, 2000; Piette & Kerr, 2006). In addition, health literacy, as measured by the short-

form Test of Functional Health Literacy in Adults (s-TOFHLA), influences medication

self-care behavior and is inversely associated with HbA1c control (Schillinger,

Grumbach, Piette, Wang, Osmond, Daher, Palacios, et al., 2002). Neighborhood factors

such as unsafe housing have been shown to present barriers to the following

recommended diabetes self-care practices known to influence outcomes: healthy dietary

11

practices, achieving recommended levels of physical activity (Dahmann, Wolch, Joassart-

Marcelli, Reynolds, & Jerrett, 2010), and smoking cessation (Cerdá, Diez-Roux,

Tchetgen, Gordon-Larsen, & Kiefe, 2010; Halonen et al., 2012).

Primary care practitioners recognize the general influence of psychosocial factors

on diabetes self-care behavior, as evidenced by their acknowledgement of the need for

better collection and use of psychosocial information (Robert Wood Johnson Foundation,

2011). The Institute of Medicine (IOM), national physician surveys, and statements from

associations representing primary care physicians, have all recommended that

psychosocial information should be included in the electronic health record. In 2014, the

IOM released a comprehensive

report in two phases detailing

the need to identify and

capture social determinitants of

health in the electronic health

record (Institute of Medicine,

2014a, 2014b). Included in this recommendation are both individual and structural

psychosocial factors (see Figure 2.2.1).

However, in addition to recommendations to improve the documentation of

psychosocial information, practitioners cite the need for better understanding of how

psychosocial factors may impede diabetes self-care behavior (Bruce et al., 2009;

Delamater, 2006; Estabrooks et al., 2012; Funnell, 2006; Ganz, 2008; Institute of

Medicine, 2008a; Krist et al., 2014b). Moreover, barriers to understanding stem from the

lack of clarity in defining the factors that may be an issue for diabetes self-care, and the

Figure 2.2.1: IOM Table of Psychosocial Factors

12

lack of tools to measure their influence (Fisher et al., 2010; Funnell, 2006). Although

tools have been created to assess psychosocial factors that influence outcomes for

conditions such as physical rehabilitation, childhood obesity, and cancer (Holland &

Bultz, 2007; Holm-Denoma, Smith, Lewinsohn, & Pettit, 2014; Wideman & Sullivan,

2012), standardized tools are not widely used for the majority of chronic conditions

encountered in outpatient care—including diabetes.

Given this lack of clarity regarding the definition and measurement of

psychosocial factors, it is important to understand how practitioners currently prioritize,

access, and use this information to support patient-centered clinical decisions, potentially

in efforts to lessen the influence of psychosocial barriers (Karazsia, Berlin, Armstrong,

Janicke, & Darling, 2014). However, little is known about what factors practitioners

consider to be relevant to their diabetes-related clinical decisions. Additionally, little is

known about how psychosocial information may influence specific diabetes care clinical

decisions, in particular, attempts to help patients address psychosocial barriers to care. As

depicted in the black box in Figure 2.2, these questions are addressed in the context of the

present research.

Clinical Practice Guidelines (CPGs) Clinical practice guidelines (CPGs) are designed to support the complex process

of clinical decision making. They represent the synthesis of best practice evidence

through, “systematically developed statements to assist practitioner and patient decisions

about appropriate health care for specific clinical circumstances” (Grossman, Field, &

Lohr, 1990; Medlock et al., 2011; Nigam, 2013; Nilasena & Lincoln, 1995; Sacks et al.,

2011; Todd, 1998). CPGs attempt to standardize clinical decisions by reducing variations

13

in these decisions. Nevertheless, CPGs are designed to only provide a framework for

decision making for practitioners; it is recognized that they do not replace clinical

judgment, practitioner-patient interaction, or the decision making process (Higgs, 2008).

As is outlined in the literature review, it is known that there are often situations in which

the diabetes care offered by clinicians deviates from clinical guidelines. While the

literature suggests that psychosocial factors may contribute to this reality, the specific

role of psychosocial information in decisions to deviate from clinical guidelines has not

yet been directly investigated. Specifically, the psychosocially-relevant situations that

prompt deviation from guidelines are not well understood. For instance, low levels of

social support are known to present barriers to following recommended medication

practices, and restrict access to healthy foods and exercise; however, the specific

situations that practitioners may consider psychosocial factors remain unclear. As shown

in the black box in Figure 2.2, this research addresses these questions.

Clinical Decision Making The literature reviewed below shows that clinical decision making is a

multifaceted process which can involve evaluating CPGs against patient-specific

information in order to individualize care—which recognizes the unique characteristics of

the patient and the patient’s situation; patient-specific information can include

psychosocial factors such as patient preferences and level of health literacy (Ceriello et

al., 2012; Hirsch et al., 2012; Radwin & Alster, 2002; Riddle & Karl, 2012). I describe

theoretical perspectives on clinical decision making, and focus my review on information

access for clinical decisions, as this area is closely associated with how psychosocial

information may be accessed, and subsequently used. Despite this prior research,

14

however, the role of psychosocial information in clinical decision making is not well

understood; this is important because psychosocial information differs from other types

of clinical information because it is typically qualitative and in narrative form. Therefore,

there is a need to understand how practitioners access and use psychosocial information

in their clinical decision making; this research addresses this question as shown in Figure

2.2. Finally, an issue related to access and use of information investigated here is the

question of barriers and facilitators to information access. This question is important to

interrogate at this time given the recent widespread adoption of digital tools such as

EHRs in outpatient care.

2.3. Psychosocial Factors - Overview and Influence on T2DM Self-Care There is no universal definition for the concept of “psychosocial factors”. Rather,

it is a general term used in various health research areas (R. J. Anderson, Freedland,

Clouse, & Lustman, 2001; Martikainen et al., 2002). For this study, I define psychosocial

factors as the psychological and social influencers of self-care behavior which impact

diabetes outcomes. These include the psychological factors—how an individual feels—

and the social factors—an individual’s social milieu—that affect the self-care behavior

known to influence health outcomes (Brotman et al., 2007; Chida & Hamer, 2008;

Institute of Medicine, 2008b; Macleod & Davey Smith, 2003; Martikainen et al., 2002;

McEwen, 1998; Plomin & Asbury, 2005; Powledge, 2011; Singh-Manoux, 2003).

Psychosocial information differs from clinical information, such as clinical

information from laboratory reports, in two ways. First, psychosocial information tends to

be qualitative in nature, and is often conveyed through stories. This information reflects

how the patient thinks about himself/herself, or their environment (Carey, 2003). As a

15

result, practitioners are inclined to interpret this information in narrative form, rather than

as specific data elements (Walsh, 2004). Subsequently, clinical decisions are based upon

how the practitioner interprets the narrative (Kay & Purves, 1996). Second, psychosocial

information tends to be situational, defined and bound by the particular patient’s

circumstances. As a result, considerable judgment is required to determine relevance and

level of influence for a particular clinical situation (MacMullin & Taylor, 1984; Taylor,

1982, 1991).

I group psychosocial factors into two categories: individual and structural.

Individual factors include health literacy and psychological issues such as an individual’s

thoughts, perceptions, attitudes and emotions. Structural factors are related to the

circumstances of an individual’s lived experience, which include their social (i.e., level of

social support) and community environment (i.e., neighborhood setting). The

neighborhood setting includes community characteristics such as housing sufficiency and

stability, level of neighborhood safety, and access to health care and healthy foods.

Culture and spirituality refer to cultural norms, faiths, beliefs and practices (Singh-

Manoux, 2003; Soto et al., 2015). Table 2.3 shows a summary of these individual and

structural psychosocial factors.

Table 2.3: Psychosocial Factors that Affect T2DM Outcomes

Individual Factors Structural Factors

SES - Social position, income, education

Social support - family support, supportive relationships, involvement in spiritual communities, and neighborhood cohesion

Activities of daily living / Responsibilities – self-care activities, family and/or work responsibilities

Culture and spirituality - cultural norms and traditions, dietary practices, faith beliefs and practices

16

Thoughts, perceptions, attitudes and emotions - Perception of discrimination, risk perception, lived experience, effort/ reward balance, perception of control/autonomy

Neighborhood setting - community violence/safety, neighborhood cohesion, housing sufficiency and stability, access to healthy foods, transportation, access to health care, access to physical activity

Health Literacy – Ability to access and understand health information that can support informing care decisions, commonly supported upon diagnosis with Diabetes Self-Management Education (DSME)

Mental Health - mood and affect, thought processes, orientation, manner and approach, alertness, appearance

2.3.1. Psychosocial Factors and Health Outcomes Psychosocial factors can exert considerable influence upon general health

outcomes. Research indicates that an individual’s living environment can have a larger

effect on their health outcomes than the specific health care services they access. For

example, neighborhood cohesion is positively associated with physical and mental health

outcomes; community violence is associated with negative physical and mental health

outcomes. Supportive social networks, which include family support and involvement in

spiritual communities, are positively associated with physical and mental health

outcomes (Robert Wood Johnson Foundation, 2011). Perceptions of discrimination, lack

of supportive relationships, and stressors are associated with negative physical and

mental health outcomes (R. J. Anderson et al., 2001; C. R. Clark et al., 2013; Delamater

et al., 2001; Egan, Tannahill, Petticrew, & Thomas, 2008; Institute of Medicine, 2008b;

Martikainen et al., 2002; Robert Wood Johnson Foundation, 2011).

Psychosocial factors can influence health outcomes in two ways, through: 1)

psychobiological processes, and 2) self-care behavior (Martikainen et al., 2002). With

17

regard to psychobiological processes, social support is associated with general

cardiovascular, endocrine, and immune system health (Uchino, Cacioppo, & Kiecolt-

Glaser, 1996). Similarly, decreasing socioeconomic status (SES) is associated with a

general increase in psychosocial stressors, which result in a range of physiological effects

associated with poor health outcomes, such as: unhealthy cholesterol profiles, behavioral

depression, and obesity (Hertzman & Siddiqi, 2009). Individuals who experience difficult

life stressors—or experience them with less capability to manage them effectively—

experience higher levels of stress, anxiety, anger, and frustration; all are associated with

poor mental and physical health (P. A. Hall & Taylor, 2009). As for diabetes in

particular, there is an extensive body of research describing the effect of psychosocial

factors like social support, SES, and stressors on physiological outcomes associated with

diabetes, such as: anxiety which is associated with poor glycemic control, and depression,

which is associated with hyperglycemia and acceleration of coronary heart disease

(Bartley, 2006; Berkman, 1995; Berkman & Kawachi, 2000; Krishnan, Cozier,

Rosenberg, & Palmer, 2010; Mirowsky & Ross, 1998).

Recommended diabetes self-care behavior can be influenced considerably by

psychosocial factors. The effect of psychosocial barriers on recommended self-care

behavior can be lessened by clinical decisions which may influence T2DM outcomes.

Therefore, for this study, my focus is on the psychosocial factors that affect self-care

behavior required for the T2DM treatment regimen, which influence health outcomes, as

depicted in Figure 2.2. This is especially important since diabetes treatment guidelines

require considerable lifestyle changes for many patients. There are seven self-care

practices vital to achieving and maintaining good T2DM health outcomes: 1) healthy

18

eating, 2) regular physical activity, 3) consistent blood glucose monitoring, 4) medication

adherence, 5) attending regular appointments, 6) healthy coping skills, and 7) risk-

reduction behaviors; all are associated with glycemic control, reduced complications, and

improved quality of life (Shrivastava, Shrivastava, & Ramasamy, 2013).

I now summarize the extensive literature describing the influence that

psychosocial factors have on diabetes self-care behavior (Cosansu & Erdogan, 2014;

Delamater et al., 2001; Duke, Colagiuri, & Colagiuri, 2009; Funnell et al., 2009;

Levinson & Roter, 1995; Siminerio, Funnell, Peyrot, & Rubin, 2007; Skovlund & Peyrot,

2005; Tapp et al., 2012; Willens, Cripps, Wilson, Wolff, & Rothman, 2011; W. Wilson et

al., 1986). In addition, I describe the influence of psychosocial factors on diabetes

outcomes, segmented by the individual and structural factors shown in Table 2.3.

Individual Psychosocial Factors Individual psychosocial factors include: SES, daily activities, thoughts

perceptions, attitudes, emotions, health literacy, and mental health. As will be shown,

these individual factors influence diabetes outcomes through their influence on self-care

behavior.

SES In a health context, socioeconomic status (SES) is defined by education, income

and health insurance payor status (Kangovi et al., 2013). Belonging to low SES groups is

associated with poor diabetes outcomes, when compared to individuals in high SES

groups. This difference is primarily due to lower adherence to one’s treatment regimen

(Jotkowitz et al., 2006). Patients from low SES groups who indicate cost as a barrier to

medication adherence experience poorer glycemic control, primarily due to financial

barriers to medication access (Piette & Kerr, 2006). Such problems may be common

19

given that diabetic patients incur higher treatment costs when compared to those with

other common chronic conditions (Dubois et al., 2000; Rogowski, Lillard, & Kington,

1997).

Activities of Daily Living / Responsibilities Following recommended self-care practices requires that the patient make a

substantial time investment, which may interfere with activities of daily living (i.e., basic

and instrumental), and family and/or work responsibilities. Diabetes self-care can restrict

the time and attention patients can devote to instrumental activities of daily living

(IADL), such as childcare and employment responsibilities, since the self-care regimen

frequently includes time-consuming activities such as: regular blood glucose monitoring

to measure fasting blood sugar (FBS) levels, medication management, physical activity,

and meal preparation (Russell, Suh, & Safford, 2005; Safford, Russell, Suh, Roman, &

Pogach, 2005). Notably, time demands tend to increase with number of health conditions,

particularly in relation to medication management (Noël, Chris Frueh, Larme, & Pugh,

2005). These increased demands are likely to be experienced by many adult T2DM

patients since a majority of them have at least one comorbid condition, while 40% have

three or more (Piette & Kerr, 2006).

Thoughts, Perceptions, Attitudes and Emotions Patient perceptions and attitudes are also associated with diabetes outcomes

through the pathway of self-care behavior. T2DM patients who report high levels of trust

in their physicians are more likely to engage in self-care known to reduce risk of disease

progression, such as smoking cessation and increased physical activity (Selby, 2010).

Also, a study showed that African American diabetes patients who perceived racial

20

discrimination in health care settings were less likely to adhere to their prescribed

medications (J. Wagner & Abbott, 2007).

Health Literacy Low health literacy is associated with poor glycemic control and high rates of the

aforementioned microvascular complication of retinopathy (Schillinger, Grumbach,

Piette, Wang, Osmond, Daher, Palacios, et al., 2002). Systematic reviews of diabetes self-

management education (DSME) programs describe how they can result in improved

outcomes, specifically for health literacy (Bielamowicz, Pope, & Rice, 2013; Duke et al.,

2009; Ellis et al., 2004; Loveman et al., 2003; Norris, Lau, Smith, Schmid, & Engelgau,

2002; Rutten, 2005; Wild et al., 2007). However, standardized diabetes education is not

uniformly accessed by patients. Low SES, low literacy, and low numeracy are access

barriers to diabetes DSME programs (D. R. Anderson & Christison-Lagay, 2008; Bowen

et al., 2013; Claydon-Platt, Manias, & Dunning, 2014; Funnell et al., 2009).

Mental Health Diabetic patients are at increased risk for mental health issues; for example,

depression is twice as common for individuals with diabetes, compared to non-diabetics

(Findley, Shen, & Sambamoorthi, 2011; Fisher, Glasgow, & Strycker, 2010; Goldney,

Phillips, Fisher, & Wilson, 2004; Heisler & Resnicow, 2008; Ward & Druss, 2015).

Strained social relationships and the emotional burden of the chronic condition contribute

to the incidence of depression, although direct causal pathways have not been identified

(Arigo, Smyth, Haggerty, & Raggio, 2014; Moulton, Pickup, & Ismail, 2015). Having a

mental health condition can present barriers to following the care regimen; diabetes

patients with depression have poorer self-care behavior when compared to T2DM

21

patients without depression (D. R. Anderson & Christison-Lagay, 2008; Piette & Kerr,

2006).

Structural Psychosocial Factors Structural psychosocial factors are related to the individual’s lived experience.

These factors include: social support, neighborhood setting, and culture and spirituality. I

now summarize the literature on the influence that these three specific structural factors

have on diabetes self-care.

Social Support Level of social support is positively associated with HbA1c at target goal (Arigo

et al., 2014; Kumari, Head, & Marmot, 2004), primarily through support with following

dietary restrictions, medication self-care, smoking cessation, and better coordination and

integration between primary care and specialty care providers (i.e., psychiatrists)

(Chwastiak et al., 2015; Heisler & Resnicow, 2008). Consistent adherence to self-care

recommendations involves not only the patient, but also their social support system

(Funnell, 2010). Family members and caregivers can provide four types of social support

that help facilitate healthy self-care practices: 1) tangible (e.g., help with activities of

daily living, transportation to medical appointments, financial assistance), 2) emotional

(e.g., encouragement, affirmation, comfort), 3) informational (e.g. reminders about

appointments, medications, dietary choices), and 4) appraisal (e.g. recognition of healthy

choices, noting clinical goals reached). High levels of social support, specifically from

spouses and other family members, are associated with increased self-care adherence to

the diabetes regimen (Iida et al., 2010; Mayberry & Osborn, 2012; Rintala et al., 2013;

Sharfi et al., 2013; Shier et al., 2013). Positive social support is a predictor for healthy

dietary behavior, foot care, blood glucose monitoring, and performing daily physical

22

activity (Tang et al., 2008; Watkins et al., 2013). Family members can also provide

support through observable actions that make following recommended self-care

behaviors easier (Mayberry & Osborn, 2012). For example, family members can

participate in physical activity and make healthy eating choices. Conversely,

unsupportive social behaviors are associated with lower adherence to the care regimen.

Although relatively rare, T2DM patients who indicate that family members disregard the

recommended diabetic diet by offering them unhealthy foods have low adherence to

dietary guidelines (Henry, Rook, Stephens, & Franks, 2013; Mayberry & Osborn, 2012;

Song et al., 2012).

Neighborhood Setting Neighborhood characteristics can exert a negative effect on diabetes outcomes

(Auchincloss et al., 2009). T2DM patients from high-poverty neighborhoods have higher

rates of: poor dietary practices, physical inactivity, and smoking (Cerdá et al., 2010;

Halonen et al., 2012). Neighborhood characteristics shown to contribute to poor diet and

low physical activity for diabetes patients include: unsafe housing (Cadzow, Vest, Craig,

Rowe, & Kahn, 2014; Jack, Jack, & Hayes, 2012; Shenassa, Stubbendick, & Brown,

2004), access barriers to healthy foods (Beaulac, Kristjansson, & Cummins, 2009) and

limited access to recreation facilities (Dahmann et al., 2010). African American patients

from low-resourced communities have expressed how neighborhood stressors present

considerable challenges to diabetes self-care behavior (Jack, Liburd, Tucker, & Cockrell,

2014; Senteio & Veinot, 2014).

Culture and Spirituality Ethnic groups share various culturally-based behaviors that influence health

outcomes. Culture is defined as the knowledge, beliefs, customs, traditions, and habits

23

shared by a group of people (Kittler, 1995). These are learned behaviors which are passed

down from generation to generation through shared norms and customs. African

American group membership has a negative influence on diabetes recommended self-care

practices, specifically dietary behavior (L. Clark, Vincent, Zimmer, & Sanchez, 2009;

Cossrow & Falkner, 2004; Scheder, 1988). This is explained by aspects of the traditional

African American diet which are inconsistent with diabetes dietary recommendations.

Hence, decreasing the sodium, fat, and cholesterol contained in traditional foods is a

substantial lifestyle adjustment for some African American T2DM patients, which

presents barriers to recommended dietary practices (Kulkarni, 2004; Senteio & Veinot,

2014). Hispanic group membership also has a negative influence on recommended

dietary behavior (Hunt, Valenzuela, & Pugh, 1998). Membership in this group is

associated with a high-fat diet in general (Daniulaityte, 2004), and a high-calorie diet

with specific dietary choices (i.e., tortillas, soda) (Barquera et al., 2008; de Alba Garcia et

al., 2007).

Spirituality can exert a positive influence on diabetes self-care in general. There is

positive association between spiritually and self-management of diabetes risk factors

(Watkins et al., 2013). For example, African American patients indicate that spirituality is

an important source of support in helping them manage the demanding diabetes self-care

regimen; they cite prayer and church attendance as sources of support (Newlin et al.,

2008; Polzer & Miles-Shandor, 2005, 2007; Quinn et al., 2001).

Gaps Addressed in This Study Most of the published literature concerning barriers to diabetes self-care behavior

is focused on the patient experience, rather than practitioners’ perspectives (Nam,

24

Chesla, Stotts, Kroon, & Janson, 2011). For example, mental health status is a

psychosocial factor known to influence diabetes self-care behavior (Beverly, Brooks,

Ritholz, Abrahamson, & Weinger, 2012; Delamater et al., 2001); there is a fairly robust

area of research examining how depression can influence diabetes self-care behavior

(Beverly et al., 2012; Chida & Hamer, 2008; Daniulaityte, 2004; Funnell, 2006;

Mayberry & Osborn, 2012). However, the literature is sparse concerning practitioners

perspectives on barriers to following diabetes self-care recommendations (Ritholz,

Beverly, Brooks, Abrahamson, & Weinger, 2014; van Dam, van der Horst, van den

Borne, Ryckman, & Crebolder, 2003). Subsequently, we do not know how treatment

decisions may be influenced by psychosocial factors. Potential differences between

patients and practitioners concerning perceptions of barriers to recommended self-care is

important as it may influence outcomes in two ways. First, this disconnect can result in

practitioners not understanding specific barriers to self-care and subsequently not

recommending available support services to address these barriers (Peyrot, Rubin, &

Siminerio, 2006). Second, these differences may limit the quality of the practitioner-

patient interaction (R. M. Anderson, Fitzgerald, Gorenflo, & Oh, 1993; R. M. Anderson

& Funnell, 2005). The quality of practitioner-patient communication is positively

associated with the following specific self-care behavior known to influence diabetes

outcomes: daily glucose monitoring (Ciechanowski, Katon, Russo, & Walker, 2001),

medication behavior, daily foot check, healthy diet, and daily exercise (Piette,

Schillinger, Potter, & Heisler, 2003).

Despite this considerable research describing the mechanisms linking specific

psychosocial factors and the quality of the practitioner-patient relationship to diabetes

25

self-care behavior known to influence diabetes outcomes, little is known of practitioners’

understanding of the relationships between psychosocial factors and diabetes self-care

behavior. Further, little is understood concerning how practitioners perceive the

comparative relevance of individual psychosocial factors and their potential impact on

self-care behavior. This research purports to fill these gaps by addressing the following

research question:

RQ1: Which psychosocial factors do practitioners perceive to be important in making, or providing input into, care decisions for adult, type 2 diabetes patients? What is their relative priority?

2.4. Clinical Practice Guidelines (CPGs) in Support of Clinical Decisions The American Diabetes Association (ADA) annually publishes standards of care

for practitioners, patients, and researchers which contain current recommendations for

diabetes care. The ADA standards represent the principal recommendations for diabetes

care in the United States (Eldor & Raz, 2009); however, several organizations develop

and distribute clinical practice guidelines (CPGs). All of these guidelines attempt to

translate current clinical insight into specific care recommendations. The ADA standards

cover type 1 and type 2, as well as gestational diabetes, in the in- and out-patient setting.

They address classification and diagnosis, testing for asymptomatic patients based on

membership in at-risk patient populations, recommendations specific for glycemic

control, setting treatment goals and thresholds, prevention and management of

complications, assessment and management of comorbidity, and objectives for improving

care. Please see Appendix D: Outline of ADA Standards of Care for an outline of the

ADA Standards of Care.

26

ADA standards carry considerable influence. They are the source of standards of

care used in many organizations and programs, including the Health Plan Employer Data

and Information Set (HEDIS) (Fonseca & Clark, 2006). The HEDIS, a program under the

NCQA, is used by more than 90 percent of health plans in the United States to measure

outcomes and performance (National Committee for Quality Assurance, 2013).

ADA Clinical Practice Guidelines The ADA maintains both process and target clinical practice guidelines (CPGs)

(American Diabetes Association, 2015). Process guidelines include frequency of

performing the HbA1c test, checking blood pressure at each clinical visit, nephropathy

screening at minimum annually, retinopathy screening upon diagnosis and annually

thereafter (less frequently following one or more normal eye exams), comprehensive foot

exam at diagnosis and at least annually thereafter, providing smoking cessation treatment

as applicable, providing access to diabetes self-management education (DMSE) and

diabetes self-management support (DSMS) upon diagnosis and as needed thereafter.

HbA1c should be tested at least twice per calendar year for patients who meet clinical

goals, and four times per year for patients whose therapy has changed or are not meeting

clinical goals.

CPGs include recommended targets for glycemic control (HbA1c < 7%),

cholesterol (LDL < 100 mg/dl), and blood pressure (<130/80 mmHG). These targets are

based on substantial empirical evidence—communicated extensively throughout the

practitioner and T2DM patient population—that shows that patients who achieve and

maintain these thresholds are at substantially decreased risk for disease-related

complications and progression. The measure of HbA1c or “A1C” is of particular note.

27

This is a laboratory blood test that measures the average level of glucose (blood sugar)

over a three month period. In general, the goal is below 7%. Younger patients who

maintain target levels recommended in the guidelines experience improved health

outcomes, improved patient satisfaction, and lower cost of care—largely achieved

through a reduction in unnecessary procedures and health crises requiring hospitalization

(D. J. Cook, Greengold, Ellrodt, & Weingarten, 1997; Gross et al., 2003; Lohr, 1994;

Schmittdiel et al., 2008).

ADA standards prioritize the treatment of hyperglycemia because the empirically-

based consensus supports that reaching and maintaining specific glycemic thresholds

reduces morbidity and complications like: retinopathy, nephropathy and neuropathy

(Nathan et al., 2009). Additional important factors include management of dyslipidemia,

hypertension, obesity and insulin resistance.

Annual updates to the ADA standards are based on new interventions—most

commonly for new medications—that result in improved outcomes. New findings from

empirical research can increase the number of therapeutic choices available to clinicians,

which also increases the uncertainty inherent in selecting the most appropriate treatment

for a particular patient.

Following from this, the ADA states that these standards are not meant to

substitute clinical judgment; “these standards are not intended to preclude clinical

judgment or more extensive evaluation and management of the patient by other

specialists as needed” (American Diabetes Association, 2015). The ADA acknowledges

that numerous risk factors contribute to the inherent complexity of diabetes care and

describe evidence supporting a range of interventions—based on specific patient

28

situation—which may improve patient outcomes. Thus, the goals and thresholds reflect

what is recommended for most patients, acknowledging that patient and practitioner

preferences, comorbidity, and patient specific factors may justify modifications to

specified thresholds.

Variety in Clinical Practice Guidelines Other than the aforementioned ADA guidelines, different organizations publish

and keep current CPGs, including: the Veterans Health Administration/United States

Department of Defense (VA/DOD) (Pogach et al., 2004; Walter, Davidowitz, Heineken,

& Covinsky, 2004), the International Diabetes Federation, the American Association of

Clinical Endocrinologists (AACE), the ADA/European Association for the Study of

Diabetes (EASD), the Canadian Diabetes Association, and the American College of

Endocrinology (ACE)/AACE Road Maps to Achieve Glycemic Control (Rodbard et al.,

2009). These organizations use CPGs to varying degrees. The multiplicity of these

guidelines introduces ambiguity into clinical decisions for diabetes care.

2.4.1. T2DM Outcomes Evaluated Against CPGs Although effective therapies have long been available to control blood glucose,

cholesterol, and pressure (Schmittdiel et al., 2008), such treatment availability has not

resulted in achievement of clinical goals for the majority of patients. A relatively low

proportion of diabetic patients meet the clinical goals specified in the ADA CPGs. In

fairly large study populations, between 21-64% of diabetic patients had HbA1c < 7%, 22-

46% had LDL cholesterol at goal, and 29-33% had blood pressure at goal; only 2-10% of

diabetic patients are at recommended levels for all three goals (Davidson, 2007, 2009;

Esposito, Chiodini, Bellastella, Maiorino, & Giugliano, 2012; Giugliano et al., 2011).

29

Diabetes outcomes vary across patient populations. Disparities persist that are

associated with a patient’s race, age, and level of education. African American and

Hispanics experience a 50% to 100% higher diabetes-related mortality rate as compared

to White Americans (Soto et al., 2015; Spencer et al., 2011; Two Feathers et al., 2005).

Conversely, cholesterol targets are achieved more frequently among African Americans

when compared to Whites or Hispanics. Shorter time since diagnosis is associated with

achieving HbA1c control. Higher level of education is also associated with achieving

blood pressure control (Resnick, Foster, Bardsley, & Ratner, 2006).

There are three levels of well-researched factors that contribute to poor diabetes

outcomes: 1) patient-specific factors, 2) practitioner level factors, and 3) healthcare

system factors. Patient-specific factors include low adherence to self-care practices

(Egginton et al., 2012; Frølich, Bellows, Nielsen, Brockhoff, & Hefford, 2010; National

Guideline Clearinghouse, 2013; Philis-Tsimikas et al., 2012; Schmittdiel et al., 2008;

Selby, 2010; Shrivastava et al., 2013; D. V. Wagner, Stoeckel, Tudor, & Harris, 2015).

There is a smaller body of research suggesting that patient-specific characteristics may

influence specific process CPGs. In one study, SES influenced whether physicians

conducted a foot exam, as patients belonging to upper status groups received foot exams

more frequently than those belonging to lower status groups (McKinlay, Piccolo, &

Marceau, 2013). Practitioner-level factors include a low proportion of patients receiving

recommended care (Davidson, 2009; McGlynn et al., 2003; Saaddine et al., 2002;

Tulloch-Reid & Williams, 2003), and complexity of care due to comorbidity (Cavanaugh,

2007; Haggerty, Roberge, Freeman, Beaulieu, & Bréton, 2012; Kahn & Anderson, 2009;

Kalyani, Saudek, Brancati, & Selvin, 2010; Piette & Kerr, 2006; Tang, Ayala,

30

Cherrington, & Rana, 2011; Versnel, Welschen, Baan, Nijpels, & Schellevis, 2011).

Specific physician characteristics, such as gender or level of experience, appear not to

influence frequency of foot exams (McKinlay et al., 2013). Healthcare system factors

include limitations of the outpatient care environment—such as insufficient time for

clinicians to learn about patients’ needs, lack of appropriate quality measures at the

individual level versus population level (Hogg & Dyke, 2011; Ovretveit, 2011;

Siminerio, Wagner, Gabbay, & Zgibor, 2009), and changing diabetes care standards

(American Diabetes Association, 2011; Cefalu & Watson, 2008; Gandara & Morton,

2011; Meddings, Kerr, Heisler, & Hofer, 2012; Morris, 2000; National Guideline

Clearinghouse, 2013).

As is the case with other chronic conditions, the principal goals of diabetes

treatment are to prevent the advancement of existing complications and to stave off the

development of future complications. Since most diabetic patients will develop and

eventually die from macrovascular complications like coronary artery disease,

cerebrovascular disease, and peripheral vascular disease, clinical care decisions target

controlling metabolic and cardiovascular risk factors for such complications. Control of

these risk factors also protects against microvascular complications like nephropathy and

retinopathy (Snow, Aronson, Hornbake, Mottur-Pilson, & Weiss, 2004).

Accordingly, to control these risk factors, CPGs aim to support clinical decisions

that will result in maintaining blood glucose, pressure control, and lipid target thresholds.

To achieve this, standards include recommending dietary practices and physical exercise,

as well as oral medications for noninsulin dependent patients. For insulin-dependent

patients, recommended treatment options include short- or immediate-acting insulin at

31

specified times during the day, in addition to oral medications for blood and cholesterol

control as appropriate, and dietary and exercise guidelines (Evert et al., 2014; Kirpitch &

Maryniuk, 2011; Todd, 1998). Smoking cessation is also strongly recommended for all

T2DM patients (Nagrebetsky, Brettell, Roberts, & Farmer, 2014).

Gaps Addressed in This Study In summary, process and target CPGs focus on specific clinical decisions, such as

frequency of performing the HbA1c test and target thresholds, while at the same time

recognizing the ambiguities and challenges of applying these guidelines to specific cases.

Despite what is known of target CPGs (i.e., HbA1c < 7), the focus of this study, little is

known as to how practitioners’ perceptions of psychosocial factors influence CPG

applicability depending upon patient circumstance. Because psychosocial factors are an

important source of ambiguity and challenge in the application of CPGs, this research

addresses the second part of the third research question:

RQ3: How do practitioners use psychosocial information? How does this information influence their specific care decisions?

2.5. Clinical Decision Making

Theoretical Perspectives on Clinical Decision Making In this section, I provide a brief overview of theoretical models of clinical

decision making focused on information use. Several theoretical frameworks describe

how clinical decisions are made. They explain clinical information usage and information

sources, attempting to depict how clinical information is stored and retrieved. They all

posit that clinical knowledge exists at various levels. Practitioners’ knowledge expands as

clinical expertise is developed and as the practitioner encounters various patient

32

situations. I briefly describe three theories for clinical information use: the physical

system theory, the theory of expert cognition, and the cognitive continuum theory—an

established theory used to examine information use for clinical decisions in the primary

care setting. In order to situate my research in the clinical decision making literature and

describe the gaps it addresses, I then discuss the limitations of these existing models.

The Physical System Theory The physical system theory posits that clinical information is stored mentally as

symbols that represent objects, events and associations between these elements (Kassirer,

Wong, & Kopelman, 2010). It is not clear how these symbols are formed nor how they

are used; however, the prevailing position is that these symbols are formed through

production rules, frames, and illness scripts. The production rule, also known as

condition-action pair, states that the symbols are stored and used in an IF-THEN format.

The IF portion represents a familiar condition—like symptoms consistent with disease

progression; the THEN portion is the action to be taken whenever the IF condition is

observed. The result could trigger a decision to treat the complication resulting from

disease progression (Kassirer et al., 2010). A frame is a decision tree, a hierarchical

structure using information to address a given clinical situation. The illness script is a

complex description of a particular clinical visit, like the primary care clinical visit

(Gigante, 2013; Monajemi, Rostami, Savaj, & Rikers, 2012). The theory assumes that the

practitioner can recall from memory individual, specific “cases,” versus general models

of disease progression. Each new case is evaluated by recalling example cases, called

“exemplars,” which inform decisions on each new case. Associations between elements

can also occur via pattern recognition. Pattern matching occurs when the physicians

33

recognizes forms of contextual information, such as patients who present with specific

symptoms, to inform clinical decisions (Monajemi, Schmidt, & Rikers, 2012).

The Theory of Expert Cognition The theory of expert cognition is a complex model that attempts to describe

information use in clinical thinking (Dreyfus, Dreyfus, & Athanasiou, 1986). The theory

describes various phases of skill acquisition in the clinical context. Clinical expertise is

primarily developed through experience with numerous patient situations, commonly

referred to as “clinical cases”. The model is appropriate for understanding the

development of medical students as they progress in their training from resident to

clinical expert as it describes the development of clinical expertise through numerous

clinical situations, informed by patient outcomes and input from colleagues.

The Cognitive Continuum Theory The cognitive continuum theory is an enduring model, frequently applied to

understanding information use for clinical decisions in the primary care context. It has

three dimensions which attempt to incorporate the different types of thinking, and how it

may be applied to different decisions: 1) structure (well-structured to ill structured), 2)

cognitive mode (intuition to analysis), and 3) time required (high to low) (Custers, 2013;

Hamm, 1988; Offredy, Kendall, & Goodman, 2008). The theory provides a general

framework to understand how a continuum of cognition and psychological processes are

associated with this expansive perspective of clinical decision making. I now describe

each of the three categories.

34

Structure Task structure is dependent upon two factors: ambiguity and complexity.

Structure has an inverse relationship with ambiguity and complexity. Highly structured

decisions are characterized by low ambiguity and complexity. Degree of ambiguity is

dependent on the availability of information which can help inform the clinical decision;

accessibility is highly dependent upon how the information is organized. Complexity is

dependent upon the number of information cues available which the practitioner believes

should be considered in care decisions. For diabetes care decisions, several cues may be

required and information may or may not be available at the specific point when choices

are presented (Thompson, Cullum, McCaughan, Sheldon, & Raynor, 2004). For example,

the common clinical decision concerning adjusting insulin treatment based on

information from an HbA1c test may depend on the availability of additional clinical and

psychosocial information such as: the patient’s response to previous dosage adjustments;

their self-care practices in administering the recommended doses; and speculation on

future patient self-care practices in administering the doses which may include

assessment of psychosocial barriers.

Cognitive Mode Cognitive mode is based on the observation that clinical decisions—particularly

those in the primary care setting—are neither purely based on intuition nor evidence-

based information. As cognitive science began describing the human mind as an

information processing system, decision scientists and cognitive psychologists developed

the notion that clinical decisions occur on a continuum from intuition to analysis; position

on the continuum defines the cognitive mode. Intuitive decisions rely on perceptions and

35

assumptions, and are less reliant on evidence-based information; analytical decisions rely

on evidence, making them less reliant on intuition (Hamm, 1988; Thompson et al., 2004).

Time Required Time required to make the decision is influenced by several clinical factors. They

include the level of practitioner collaboration with the patient for a clinical decision, and

the information exchanged between practitioners. The level of collaboration with the

patient can vary based on caregiver or support network involvement in the decision.

Information exchange can vary based upon the number of practitioners that comprise the

care team of a particular practice.

Limitations of Existing Models These models do not entirely describe the complex, contextual relationship

between the various types of clinical decisions and specific psychosocial information

processing and use. For example, several additional contextual factors influence

information use, such as: the nature of the clinical decision; patient-specific psychosocial

information; the number of decisions made and the timeframe in which to make them; the

amount and priority of contrary clinical information available to the practitioner; and how

the information is organized and displayed (Lomas & Haynes, 1988; Rycroft-Malone et

al., 2004; Thompson et al., 2004; Weiner et al., 2010).

Gaps Addressed in This Study These limitations illuminate the gaps that this work can fill. As I described in

section 2.3, there are various psychosocial factors that define the patient situation,

particularly as it pertains to diabetes self-care behavior. There is a need for a model to

describe how psychosocial information is used for patient-centered diabetes care

36

decisions. To address this need, I investigate the following research question, the first

part of the third research question:

RQ3: How do practitioners use psychosocial information?

2.5.1. Clinical Information Access for Clinical Decision Making Although there is a vast body of literature on clinical information access for

decision making previous work does not discuss the unique case of psychosocial

information. I now summarize this relevant prior literature of the sources of clinical

information which is not exclusively psychosocial, and how that information is accessed.

I briefly outline the literature on information seeking and conclude this section with an

overview of the literature concerning unmet information needs.

Overview Clinical information typically focuses on the physical manifestations of patient

health, and may include subjective information such as the patient’s description of their

complaint and objective information such as laboratory or other diagnostic tests and

findings from a physical examination. Clinical information gathering can vary widely

within the primary care setting; it is not conducted according to a specific pattern during

the clinical visit. For experienced practitioners, typically information is initially gathered

from the medical record—which includes laboratory results—then the physical exam,

then from the patient (Kassirer et al., 2010). But clinical information also may be first

presented during the patient exam (i.e., physical appearance, gait), then from the

laboratory results (i.e., HbA1c, LDL), then from the consultation (i.e., pain, vision

issues). The use of other sources, such as peers, may also be woven into practitioners’

information seeking behavior.

37

Electronic Health Record (EHR) The electronic health record (EHR) is a key source of information used in clinical

decision making. In the outpatient setting, physicians order diagnostic tests and use

information from their results, which are captured and displayed in the medical record, to

inform clinical decisions (Veinot, Zheng, Lowery, Souden, & Keith, 2010). Primary care

physicians in the outpatient setting seek information most frequently about a specific

clinical situation, particular to the specific patient’s situation (Covell, Uman, & Manning,

1985; Del Fiol, Workman, & Gorman, 2014).

The EHR can be both a barrier and a facilitator of information exchange in the

outpatient care setting (Shachak & Reis, 2009). EHR tools affect care delivery as

facilitators or barriers to the quality of communication during clinical consultation,

subsequently impacting health outcomes (Asan, Montague, & Xu, 2012; Haskard,

Williams, & DiMatteo, 2009; Veinot et al., 2010; Ventres et al., 2006). EHR use can

present barriers to effective practitioner-patient communication because it can hinder

dialogue and reduce focus on the patient, key aspects of patient communication (Makoul,

Curry, & Tang, 2001; Margalit, Roter, Dunevant, Larson, & Reis, 2006; Ventres et al.,

2006). Yet practitioner use of the EHR can also facilitate communication through

improved access to patient information—particularly concerning medications—expressly

important in care for patients with multiple prescriptions (Arar, Wen, McGrath,

Steinbach, & Pugh, 2005; Ash et al., 2007; Doyle et al., 2012; Lelievre & Schultz, 2010;

Shield et al., 2010). EHR use in the exam room can have positive effects on patient

satisfaction, a key health outcome (Irani, Middleton, Marfatia, Omana, & D'Amico,

2009).

38

Computerized Decision Support Systems (CDSS) In general, Clinical Decision Support Systems (CDSS) aim to support clinical

decision making by standardizing them. Standard decisions couple the individual patient

circumstances with the evidence-based research (Berg, 1997; Dowding, 2008; Tierney,

Overhage, & McDonald, 1996). These systems vary in their features and capabilities. In

general, they utilize rules and/or statistical formulas to make expert knowledge accessible

to the practitioner at the point of the care. They can be implemented as ‘passive’ systems

that provide information only when requested by the practitioner, or as ‘active’ systems

that automatically provide information. CDSS use alerts and reminders based on practice

guidelines. These electronic reminders are integrated into the medical record to provide

practitioners with recommendations for preventive services and clinical targets (Nemeth,

Ornstein, Jenkins, Wessell, & Nietert, 2012; Strayer, Shaughnessy, Yew, Stephens, &

Slawson, 2010).

Peers In the primary care setting, nurses tend to perceive colleagues as more useful

sources of supplemental information (i.e., not contained in the EHR) for clinical

decisions rather than published research, including peer-reviewed journals, websites, and

pamphlets. Also, physicians consult with other physicians and health professionals most

frequently. Physician peers are a more common source of information regardless of type

of outpatient care practice (Covell et al., 1985; Kahane, Stutz, & Aliarzadeh, 2011).

Gaps Addressed in This Study Although the above literature highlights the importance of several sources of

clinical information for primary care practitioners, research has not yet addressed the

39

specific case of psychosocial information. However, this is important since psychosocial

information, as described previously, is unique in that it tends to be qualitative and

situation-dependent. Prior research in an inpatient setting showed that psychosocial

information has a tendency to go undocumented (Zhou, Ackerman, & Zheng, 2009)—

suggesting that access to such information may be different than access to clinical

information currently collected. Moreover, prior research has detailed the sources of

information that clinicians use, with less attention accorded to the methods by which

practitioners acquire information from such sources. A clearer understanding of how

practitioners currently access pertinent psychosocial information will help support the use

of psychosocial information in outpatient chronic care. By enhancing practitioners’

understanding of psychosocial factors and specifically how psychosocial information is

different from clinical information currently collected and used, will help achieve the

expansive goal of better collection and use of that information. To address these areas of

uncertainty, this study investigates the following research question:

RQ2: How do practitioners access psychosocial information?

2.6. Gaps in Understanding Psychosocial Information Access and Use I conclude the literature review with an overview of specific gaps in

understanding three potential areas of influence on psychosocial information access and

use: situations, practitioner characteristics, and facilitators and barriers. I use this

discussion of gaps to situate my remaining research questions in the literature. Although

the literature I summarized concerning clinical decision making and CPGs is quite

extensive concerning general clinical information use, the literature is comparatively

sparse concerning the influence of psychosocial information on clinical decisions.

40

Practitioner Perspectives on Psychosocial Factors Practitioners do not necessarily feel equipped to address psychosocial needs

(Institute of Medicine, 2008b). In a national survey of primary care physicians, only 20%

of respondents indicated that they felt confident in their ability to address the social needs

of their patients which affect health outcomes. Also, 75% indicated that the health care

system should support providing access to support services for patients—if they

determine that lack of social support presents barriers to care. Physicians indicated that if

they could write prescriptions for social support, as they do for medications, they would

write prescriptions to support: transportation assistance, access to healthy foods, and

exercise programs (Robert Wood Johnson Foundation, 2011).

Psychosocial Information Currently Used Practitioners acknowledge the importance of improving the collection and use of

psychosocial information (Robert Wood Johnson Foundation, 2011), but neither prompts

for consideration, nor sources of psychosocial information, are consistently described in

the literature for diabetes clinical decisions. Patient-specific situations appear to prompt

practitioner consideration of psychosocial factors (Weiner, 2004; Weiner et al., 2010).

Weiner has suggested a three-step framework for practitioners to avoid “contextual

errors” in identifying these situations which stem from improperly incorporating the

patient’s situation into the clinical decision. The three suggested steps are: 1) effectively

discern contextual cues, rooted in psychosocial factors (i.e., financial strain, social

support, culture, mental health status), 2) be aware of one’s own perceptions that may

bias recognition of the cues, and 3) take a systematic approach to confirm results of their

clinical decisions from multiple sources.

41

Recent published literature on “contextual errors” describes how practitioners use

the medical record to identify when patients do not achieve glycemic control. In this

situation, practitioners may perceive that psychosocial issues are present, prompting

consideration for psychosocial information such as: transportation needs, SES stressors,

and demands from daily activities (Weiner et al., 2014). Practitioners may also use the

medical record to assess medication adherence, reviewing patients’ refill patterns to help

determine medication self-care behavior (Veinot et al., 2010).

The literature has offered suggestions for reducing “contextual errors” in

assessing psychosocial factors that may help define the patient’s situation, but the

literature does not describe when practitioners currently access, and use psychosocial

information to make, or to influence, their diabetes clinical care decisions in the

outpatient setting. Therefore, the research presented here investigates the following

research question:

RQ4: In which situations are psychosocial factors considered?

Practitioner Role and Use of Psychosocial Information The type of practitioner role appears to influence use of psychosocial information.

Nurses are more likely than physicians to collect and use psychosocial information. They

more frequently believe that psychosocial factors influence self-care practices and

attempt to provide, or make referrals for, support for addressing psychosocial barriers

(Funnell, 2006; Peyrot et al., 2006). Also, nurse practitioners report conferring with nurse

practitioner peers—rather than reference manuals, or physicians—concerning ways to

address psychosocial barriers (Rasch & Cogdill, 1999). Since practitioner role appears to

have an influence on perception of relevance of psychosocial factors, and may influence

42

resulting clinical decisions, I seek to investigate any associations between practitioner

role and psychosocial information use. Insights on differences between practitioner roles

will help inform the development of capabilities to provide access to psychosocial

information when it is most needed. This is the focus of the fifth of my six research

questions:

RQ5: What practitioner characteristics (i.e. role, age) are associated with their use of psychosocial information?

Current Tools in Support of Psychosocial Information Use

Practitioners across specialties acknowledge the need for better collection,

analysis, and use of psychosocial information. There are various recommendations to

expand current capabilities of the electronic health record (AHIMA, 2014; Chunchu,

Mauksch, Charles, Ross, & Pauwels, 2012; Institute of Medicine, 2014a, 2014b; Krist et

al., 2014b; Pearson, Brownstein, & Brownstein, 2011; Zhou, Ackerman, & Zheng, 2010).

The EHR should include capabilities to access and use psychosocial information, which

the patient should be closely involved in providing. Recommendations to enhance EHR

capabilities to document and use psychosocial information such as a patient’s emotional

health and stressors. These factors are important to help inform clinical decisions

(Estabrooks et al., 2012; Glasgow, Kaplan, Ockene, Fisher, & Emmons, 2012).

In a national survey of 1,000, primary care physicians identified access to three

psychosocial-related needs critical to influencing patient health outcomes: fitness

programs, healthy foods, and transportation. A large majority of respondents (85%)

indicated that they do not feel confident in currently available tools to assess and address

patients’ social needs which influence health outcomes, such as access to: safe housing,

transportation, and healthy foods. For physicians who practice in low-SES areas, 95%

43

indicated they do not feel confident in addressing these needs (Robert Wood Johnson

Foundation, 2011).

Electronic health records could better support primary care by documenting social

needs. The current, common practice of collecting information on family social history is

not sufficient (Klinkman & van Weel, 2011). A consensus statement from the American

Academy of Family Physicians (AAFP), American Academy of Pediatrics (AAP),

American Board of Family Medicine (ABFM), and North American Primary Care

Research Group outlined patient-specific information needs in the primary care setting,

which include psychosocial information (i.e., social support, occupational information)

(Krist et al., 2014a). They acknowledge that current information collection practices are

not suitable for a robust understanding of patients’ situations, which is necessary for

providing patient-centered care (Evans & Trotter, 2009).

Improving the capture and use of psychosocial information would help

practitioners assess and respond to barriers to healthy self-care practices. First,

practitioners must have the tools to support the assessment of self-care behavior and the

factors that may impede motivation and capabilities. Second, they must work with the

patient, and caregivers as appropriate, to collaborate on a plan of action that is informed

by the patient’s situation. Third, they must have the tools to create and monitor clinical

and psychosocial health in order to provide ongoing assessment and support (Heisler &

Resnicow, 2008).

Despite this body of extant research, prior research is disjointed with respect to

understanding the full range of barriers and facilitators of the use of psychosocial

information in outpatient care. Moreover, prior work has scarcely considered issues of

44

current modes of documentation of such information. The final research question

purports to fill these gaps:

RQ6: What are the barriers and facilitators to acquiring and using psychosocial information? How effectively do current tools (templates, data fields, free text) support the storage and retrieval of psychosocial information?

I now revisit the initial conceptual framework introduced in section 2.2. In the

updated conceptual framework (see Figure 2.6) I have replaced the black box in the

initial conceptual framework (see Figure 2.2) with the specific knowledge gaps in the

literature, according to each of the six research questions which represent the focus of

this study.

* - PFs – Psychosocial Factors * - PI – Psychosocial Information * - CPGs – Clinical Practice Guidelines ** - Clinical Decision Marking mediates influence of PFs on self-care (i.e., when recommending resources for food, housing, etc.)

Figure 2.6: Conceptual Framework of Psychosocial Factors in Outpatient Diabetes Care

Summary Practitioners must have consistent access to psychosocial information to

understand if psychosocial factors may be influencing a patient’s diabetes self-care

45

behavior. In order to do so, the practitioner must understand what psychosocial factors

influence self-care, how to access them, and how to use psychosocial information to

inform diabetes clinical decisions. Further, they must determine when patients’

circumstances warrant consideration.

Little is known about practitioner perceptions of the importance of psychosocial

information, nor how it is accessed and used at the point of care. Further, little is

understood about the sources of this psychosocial information, and the barriers and

facilitators to its use. This study attempts to fill these gaps in order to better support

clinical decision making, and ultimately improve outcomes for outpatient diabetes care.

46

CHAPTER 3

RESEARCH DESIGN

We don’t receive wisdom; we must discover it for ourselves after a journey that no one can take for us or spare us.

― Marcel Proust (1871–1922)

This study followed a mixed methods design. Combining qualitative and

quantitative methods is valuable, as it leverages the strengths of both approaches (Curry,

Nembhard, & Bradley, 2009). I selected the exploratory sequential design, in which

qualitative data is collected first, and emphasized throughout the study (Creswell, 2013).

In the first phase, I used the grounded theory approach in collecting and analyzing the

data from seventeen in-depth, semi-structured interviews with physicians, conducted in

five states (Charmaz, 2006). The results of this analysis informed the design of the online

survey. The online survey enabled me to explore if data gathered from physician

interviews can be generalized to a larger sample of practitioners, and to assess

relationships between variables. The specifics of the research design are outlined below.

3.1. Research Plan / Approach As depicted in Figure 3.1, I collected qualitative data in an initial phase (Creswell,

2013; Fetters, Curry, & Creswell, 2013). The qualitative data served to identify which

categories are present in the data; I conducted the qualitative phase once I reached

“saturation”, a situation in which analysis ceases to generate new categories (Corbin &

Strauss, 2007; Guest, Bunce, & Johnson, 2006; Ogedegbe, Mancuso, Allegrante, &

47

Charlson, 2003). I reached saturation after seventeen in-depth, semi-structured interviews

with physicians.

Figure 3.1: Exploratory Sequential Study Design (based on Creswell, 2013, p. 220)

* - CAPS are used to show emphasis on qualitative data

The qualitative data collection and analysis was followed by quantitative data

collection and analysis, which served to assess the prevalence of the categories identified

in the interview analysis, and to assess relationships between variables. To assess

generalizability, the online survey sample is comprised of 229 participants in various

practitioner roles, including physicians (MD and DO), physician assistants (PA), nurse

practitioners (NP), registered nurses (RN), registered dieticians (RD), clinical

pharmacists, and other practitioner roles.

3.2. Timeline I began the investigation in February, 2014 by conducting the initial interviews

for the pilot data collection and analysis, as shown in Figure 3.2. I continued recruiting

study participants and conducting the interviews through the summer of 2014. I

performed the interview data analysis, which helped inform my follow-up probes. I

constructed and built the survey instrument in fall of 2014, and launched it in November.

I analyzed the survey responses in late 2014, through the spring 2015. I began writing the

findings in the late spring, early summer of 2015.

48

Figure 3.2: Study Timeline

3.3. Study Population, Sampling Plan & Participant Recruiting Methods The target population for this study was practitioners who self-report experience

treating adult, T2DM patients in the outpatient setting. Eligibility did not depend upon

timing for providing this care. For example, a participant may be making, or providing

input into, these clinical care decisions in their current role, or they may have done so

during their past clinical experience. My population of interest is practitioners from

various parts of the United States.

Study Participants I chose to focus on physicians in the beginning of the study, given their central

role in clinical decision making. My interview participants are physicians who indicate

they have experience treating adult, T2DM patients (see Appendix E: Interview

Participants). I included both medical doctors (MDs) and doctors of osteopathic

medicine (DOs), including residents and fellows, with diversity of experience in

providing care for T2DM patients, according to the sampling strategy detailed below. I

49

conducted these interviews on site at the physicians’ preferred location. The interviews

took place in five states: California, Indiana, Michigan, Rhode Island, and Texas.

As I continued the study, I decided to include additional practitioner roles in the

quantitative step in order to obtain perspectives from a wider group. The online survey

was distributed to a larger, more diverse sample of practitioners which included:

physicians (MD and DO), resident physicians, physician assistants (PA), nurse

practitioners (NP), registered nurses (RN), registered dieticians (RD), and clinical

pharmacists. In order to sample a diversity of practitioners, these participants were

recruited from three sources: 1) Genesis Physicians Group, 2) North Texas Nurse

Practitioners, and 3) Michigan Association of Diabetes Educators.

Genesis Physicians Group Genesis Physicians Group (Genesis) is based in North Texas. Genesis is North

Texas’ largest independent practice association (IPA). Genesis has 1,400 physician

members, 450 who are primary care physicians. Genesis supports its physician members

by helping them manage their practices, focusing on enhancing economic value (Genesis

Physicians Group, 2015).

North Texas Nurse Practitioners The North Texas Nurse Practitioners (NTNP) is a 700 member professional

organization in the North Texas region. It is recognized as an affiliate member of the

Texas Nurse Practitioners (TNP) and the American Association of Nurse Practitioners

(AANP). The NTNP is a non-profit organization that provides local networking

opportunities, and encourages the educational and professional advancement of its

members. It is active in precepting and mentoring activities, and especially focuses on

50

informing the general public of the nurse practitioner profession (North Texas Nurse

Practitioners, 2015).

Michigan Association of Diabetes Educators The Michigan Association of Diabetes Educators is an affiliate of the American

Association of Diabetes Educators (AADE). The affiliate has approximately 400

members across Michigan. The AADE network helps facilitate communication

throughout its affiliates, at the state and local level. The AADE also has communities of

interest (COI) based on professional practice areas (American Association of Diabetes

Educators, 2015).

Sampling Plan and Recruitment In keeping with the principles of grounded theory (Bryant & Charmaz, 2007), my

sampling and recruitment was driven by the phenomena I was seeking to investigate. I

targeted practitioners with experience treating complex, T2DM patients in outpatient care

settings. Complexity is determined by the presence of comorbid conditions, and low-

resourced individuals – especially from ethnic minority groups.

Interview Sampling My sampling approach for interview participants followed a multi-phased,

grounded theory sampling strategy. I used convenience sampling during the pilot phase

of the research—sampling the most accessible study participants—for the initial six study

participants (Marshall, 1996). These individuals were recruited through my personal and

professional network. The pilot phase study participants represented various levels

clinical experience (2nd year resident to 19 years), geography (central to eastern United

States), and training (family medicine, internal medicine, endocrinology). My

preliminary analysis of the pilot interview data revealed that physicians consider the

51

following psychosocial factors in their clinical decision making process: socio-economic

status (SES), culture, immigrant status, and comorbidity (complexity based on multiple

chronic conditions, including mental health conditions such as depression, and that the

EHR played an important role in the use of psychosocial information. According to

grounded theory, I then used purposeful sampling based upon how the interview

participants described psychosocial factors and their use. In particular, sampling sought

variations in levels of experience to understand how the advent of EHR tools impacted

participants’ ability to document and use psychosocial information, a key theme that

emerged in the early physician interviews. Last, I used theoretical sampling according to

my emerging theory (Morse, 2010). This included sampling based on experience with

patients who are poorly controlled, from low-resourced areas, and have a low ability to

pay for treatment.

Survey Sampling Survey participants are individuals who make, or provide input into, diabetes care

decisions in the outpatient care setting. The population of interest includes practitioners

across the United States who self-report either making T2DM clinical care decisions, or

providing input into, these clinical care decisions that are made by others (e.g., a

registered nurse who provides input into a physicians’ referral or prescribing decisions).

To increase the odds of finding practitioners with relevant expertise, I sampled primary

care physicians and members of a professional organization of diabetes educators, many

of whom are nurses. I also included a nurse practitioner professional organization to

ensure perspectives from diverse practitioner roles.

Response rates have steadily declined from 2000 – 2012 for large-scale surveys

conducted with various medical practitioners (Klabunde, Willis, & Casalino, 2013;

52

Wiebe, Kaczorowski, & MacKay, 2012). Therefore, I referred to the extensive literature

on maximizing response rates for healthcare professionals, specifically physicians (Cho,

Johnson, & VanGeest, 2013; Grava-Gubins & Scott, 2008; Hawley, Cook, & Jensen-

Doss, 2009; Jansen et al., 2005; Thorpe et al., 2009; VanGeest, Johnson, & Welch, 2007;

P. M. Wilson, Petticrew, Calnan, & Nazareth, 2010).

This literature revealed the potential value of identifying peers to help promote

survey engagement; this approach has been used in administering online surveys to

healthcare practitioners (Grava-Gubins & Scott, 2008). Therefore, I recruited survey

“champions” within each of the three groups to support survey distribution and maximize

response rates. Survey champions are influential peers, and recognized leaders in the

survey sample. My specific approaches for subsample are identified below.

Survey Sampling – Genesis Physicians Group The Chief Executive Officer (CEO) of Genesis Physicians Group served as

survey champion for the primary care physicians survey sample. Working closely with

him, I developed a survey recruitment strategy integrated with the launch of their

Accountable Care Organization (ACO) initiative, which includes segmenting their 113

primary care physicians into six groups, called PODS. Each of these six PODS are led by

Associate Medical Directors (AMDs), who serve various roles, including communicating

with their PODS important information about the progress of the ACO initiative. The

online survey represented the first major communication from the AMDs to their PODS.

The Genesis CEO served as the champion for informing the AMDs about the survey, and

distributing reminders to their PODS to complete the survey.

53

Survey Sampling – North Texas Nurse Practitioners The president of the North Texas Nurse Practitioners (NTNP) professional

organization served as the survey champion for this group. I worked with her to

familiarize her with the study objectives, and articulate specifically what I was seeking

from the survey sample. The president provided me with their 257 membership

distribution list, which included email addresses. This was essential to tracking individual

responses.

Survey Sampling – Michigan Association of Diabetes Educators An eminent member served as the champion for the Michigan Association of

Diabetes Educators. The membership list was not available to me or the champion.

However, the survey champion sent the survey email, which included a link to the

survey, to their Listserv, an electronic email list used for communication to all 399

members. Please see Appendix F: Emails Used for Survey Distribution for the emails

used for survey distribution.

3.4. Data Collection The sample for the qualitative (interview) data was distinct from the sample for

the quantitative (survey) data, and the interview sample population is smaller than the

survey sample. I purposively minimized the time between the interview data collection

and analysis, and the survey construction and distribution (Harris & Brown, 2010).

3.4.1. In-Depth, Semi-Structured Interviews I conducted individual, in-depth semi-structured interviews (Lillrank, 2012;

O'Reilly, 2012), in a private location. The interviews generally lasted between 45 minutes

to one hour. The average length of the interviews was 58 minutes, 33 seconds. The

shortest interview was 34 minutes, 50 seconds (P03); the longest interview was one hour,

54

20 minutes, and 49 seconds (P15). Informed by the extant literature, I used open ended,

main questions and follow-up probes (see Appendix G: Semi-Structured Interview

Guide). I selected this approach because it is well-suited for collecting information from

participants about a specific topic (Britten, 1995; Hesse-Biber & Leavy, 2010).

Prior to the interview, I informed each participant of the purpose of the study, and

obtained their informed consent. I kept a separate file containing identifying information

in a secure location on a password-protected computer using my University of Michigan

Google drive account.

Insights gleaned from the initial interview participants helped guide my probes in

subsequent interviews. I continued to conduct interviews until my analysis reached

saturation. All interviews were conducted face-to-face. The participant determined the

location; I traveled to their preferred location.

All interviews were transcribed verbatim. I used Scribie, an external transcription

service. All interview transcripts were de-identified, and interview participants were

given an anonymous ID number (i.e., P01).

3.4.2. Survey Design The survey is a common research tool used across numerous areas of research.

There is a common process to its design. Groves et al. (2011) outline the process, which I

have adapted and depicted in Figure 3.4.2. In this section, I will describe how I

constructed the survey, and validated it using cognitive interviews. The recruitment and

sampling plan have been described in section 3.3. I describe how I analyzed the survey

data in section 3.5.

55

Figure 3.4.2: Survey Design Process (based on Groves et al., 2011, p. 47)

I constructed the survey instrument based on the categories identified in the

interview data. Specifically, the survey asked about what psychosocial factors

practitioners use, their relative priority, what decisions are influenced by these factors,

and what triggers their consideration. I also asked respondents to identify their

information sources for psychosocial information, their practice setting, and what clinical

decisions they personally make, or provide input into. This was important because nurse

practitioners, physician assistants, and clinical pharmacists may not make, or provide

input into the same clinical decisions as physicians across practice settings. Additionally,

allied health professionals may find themselves providing input into, but not themselves

ultimately making, decisions that lie within the scope of practice for practitioners with

prescribing authority. Further demographic details, such as years of practice, were also

gathered to characterize the survey sample.

In constructing the survey questions, I followed proven principles of writing good

questions based on the survey methodology literature (Passmore, Doobie, Parchman, &

Tysinger, 2002). Key principles include:

• Use simple words, that all respondents will understand • Use memory cues to improve recall • With closed questions, include all reasonable, possible responses

For additional detail please see Appendix J: Principles of Survey Instrument Design.

56

Validating the Survey Instrument – Cognitive Interviews I validated the survey instrument prior to finalizing and distributing the survey

using the cognitive interview approach. Cognitive interviews have been used to pretest

health surveys prior to distribution (Drennan, 2003) and I believed the approach was

suitable given the complexity of the research topic.

Therefore, I used cognitive interviewing to pretest the questions in order to detect

any issues with clarity or meaning (Beatty & Willis, 2007; Sherman et al., 2014; Willis,

2005). Cognitive interviewing entails administering the survey to a portion of the sample

population, while collecting additional verbal information about the survey instrument

(Beatty, 2003). I ensured quality of responses by confirming that the questions were well

understood and easy to answer (Groves et al., 2011; Jansen et al., 2005). I also assessed

how well the answers corresponded to what I intended to measure. I confirmed that all

questions followed three standards: 1) content standards—the questions ask about the

right elements, 2) cognitive standards—the respondents understand the questions and

have the information required to answer them, and 3) usability standards—respondents

can complete the survey easily (Groves et al., 2011; Squires et al., 2013).

I conducted seventeen cognitive interviews from September 25, 2014 through

October 15, 2014 (see Table 3.4.2). Each of the following clinician roles were

represented in the cognitive interviews: physician, physician assistant, nurse practitioner,

registered nurse, and clinical pharmacist. The interviews lasted an average of 33 minutes,

32 seconds; the longest lasted one hour, 40 minutes; the shortest lasted 19 minutes.

During the cognitive interviews, I gained feedback based on what the respondents

thought as they reviewed the questions. I made the specific adjustments to the survey

based on feedback in the following areas: survey look and feel (e.g., finalized groupings

57

of clinical decisions), clarity of wording of questions and answer choices (e.g., added

context on circumstances when psychosocial factors are important), and specific content

changes (e.g., added “selecting generic vs. brand” to medications decisions answer

choices). I recorded these interviews and referred back to them as needed prior to

finalizing the instrument. Please see Appendix H: Survey Instrument.

Table 3.4.2 – Cognitive Interview Participants

In Person Telephone TOTAL Clinical Role

Physician 2 1 3 Physician Assistant 0 3 3 Nurse Practitioners 1 4 5 Registered Nurse 0 4 4 Clinical Pharmacist 1 1 2

TOTAL 4 13 17

3.4.3. Survey Distribution and Administration I used Qualtrics software to create the instrument and distribute the online survey

via email, and collect the data (see Table 3.4.3). I used the University of Michigan email

survey for all correspondence with the survey participants. The email system is protected

with password access. I assumed that all study participants were familiar with standard

email security protocols. I introduced the survey in the survey email, and I informed the

participant of the purpose of the study and obtained their informed consent. As I did with

the interview data, I kept a separate file containing identifying information for survey

participants in a secure location on a password-protected computer using my University

of Michigan Google drive account.

I received my first survey response on November 20, 2014, and my last response

on May 5, 2015. Since the survey was distributed via email, I identified the purpose of

the study, why they were being contacted, identified the survey “champion” for the

58

particular subsample, and gave the recipients the option of removing their email address

from the distribution. Please see Appendix F: Emails Used for Survey Distribution for

emails used to distribute the survey instrument.

Table 3.4.3 – Survey Distributions

Sent by Sent From Date / Time (ET)

Genesis Physicians

Initial Email Sent to (6) AMDs Champion Qualtrics 2.3.2015

3:55pm

Email Sent to PODs AMDs Qualtrics 2.5.2015 10:00am

Reminder Fax Champion Genesis Fax 3.6.2015

Reminder Email AMDs Qualtrics 3.6.2015 6:00pm

Reminder Email Champion Qualtrics 3.11.2015 10:00am

Reminder Letter Champion Postal Mail - Dallas 4.7.2015 North Texas Nurse Practitioners NTNP

Initial Email Champion Qualtrics 11.20.2014 10:00am

Reminder Email Champion Qualtrics 12.4.2014 10:23am

Reminder Email Champion Qualtrics 12.18.2014 12:16pm

Reminder Email Champion Qualtrics 1.5.2015 10:00am

Reminder Email Champion Qualtrics 1.20.2015 1:33pm

Michigan Diabetes Educators

Initial Email Champion To Listserv 12.2.2014 Reminder Email Champion To Listserv 12.18.2014 Reminder Email Champion To Listserv 1.5.2015 Reminder Email Champion To Listserv 1.23.2015

Note: reminders were only sent to non-respondents for Genesis and NTNP.

3.5. Data Analysis Separate data analyses were conducted for the two phases of the study; these are

outlined below.

59

3.5.1. Interview Analysis I used the grounded theory approach to code, summarize, and condense the data

(Charmaz, 2006). I selected this approach because it is well suited for analysis of the

interview data (Charmaz, 2006), and because it is conducive to understanding how a

process works (Creswell, 2006). I applied this approach to understanding the

psychosocial and situational factors used in the decision process in the course of patient-

centered diabetes care decisions.

I coded the interview transcripts using NVivo version 10.0 qualitative data

analysis software. My objective was to find repetitive patterns, focused on the

psychosocial factors, the triggers of consideration, and how they are used. I coded line-

by-line to segment the data and link the patterns I identified. Coding is a cyclical process,

enabling the refinement and highlighting of patterns to generate categories and concepts.

I followed a coding process consistent with grounded theory. I did line-by-line coding in

the initial coding phase, using in vivo codes to capture the physicians’ meanings (Glaser,

1978). I summarized basic topics in a word or short phrase, using gerunds to help me

investigate processes (Miles, Huberman, & Saldaña, 2014; Saldaña, 2009). In the second

coding cycle, I used axial coding to define conditions and actions (Corbin & Strauss,

2007). Consistent with grounded theory, in this second cycle, I developed categories for

the first cycle codes, an approach well suited to apply meaning to the data (Miles et al.,

2014; Saldaña, 2009).

I calculated interrater reliability to determine the extent to which multiple coders

agree. I had a second coder code selected interview transcripts; 24% of the interview

transcripts were randomly selected by the second coder, four of the total seventeen. The

interrater reliability (IRR) between the second coder and I was 98.436% for all codes

60

used in the interview data analysis. This is above the generally recommended 90%

threshold (Di lorio, 2006).

I used memos and diagrams to describe the emergent concepts. I reviewed the

transcripts for key concepts, using the participants’ own words. Please see Appendix I:

Code Book for Physician Interviews for the codebook used for the physician interview

analysis.

The cognitive map is an established tool to depict how an individual thinks about

a particular process (Miles et al., 2014). Since there is no tool that I am aware of to depict

how a group of individuals think about a particular process, I used the cognitive map to

depict what I heard from the physician interview participants as they described how they

used psychosocial information in the course of making clinical decisions.

I constructed the cognitive map by completing two separate steps. First, I divided

key concepts into two separate categories: 1) process concepts – how decisions are made,

and 2) content concepts – what information is used to make the decisions (Miles et al.,

2014). Process categories included context on clinical decisions, such as how clinical

targets are established using psychosocial information. Content categories included what

factors represented the substance of the topic, such as what influenced their consideration

of clinical practice guidelines and psychosocial factors. Each of these categories

represented content that was present across multiple interviews. Second, I used displays

to separate content and process terms, and grouped concepts that seemed to belong

together. I did so by examining relationships, if any, that existed between concepts. By

analyzing relationships on a large display I was able to create groups of concepts, and

links between groups. I used the memos I wrote from my analysis of the interview data to

61

help identify themes. This formed the basis of the initial cognitive map, which I revised

over several iterations to create the final map depicted in chapter five.

3.5.2. Survey Data Analysis Analyses of the survey data began with data processing and identifying invalid

data. I corrected invalid entries for the five physicians who indicated non-physician roles.

I clarified their credentials with the marking director, the point of contact in the

organization. She located their credentials and training using their internal personnel

records. Once confirmed, I corrected their survey entries to “MD” or “DO”. This

approach improved the quality of the data, and confirmed that the data contained the

information intended in the design (Groves et al., 2011). Also, I confirmed quality using

the following: range edits, balance edits, and consistency edits. Range edits confirm that

data entered aligns with practical ranges. Balance edits are associated with percentages

and outliers. Consistency edits check for uniformity.

Common approaches to handling missing data are listwise deletion and casewise

deletion. Listwise deletion is omitting respondent entries that contain missing data.

Casewise deletion involves defining a case, stipulating that if any specified data elements

are missing, the entire response is deleted. I did not use listwise nor casewise deletion.

Given the time required to complete the survey, I expected, and received, incomplete

responses. I did not adjust for non-response bias, given that I did not know any

characteristics of the respondents, aside from their membership in one of the three

organizations represented in the survey sample. I included the responses submitted up to

the point that respondents stopped entering responses (Little & Rubin, 2002).

62

Statistical Analysis I initially examined survey data using descriptive statistics (Bartholomew, Steele,

Galbraith, & Moustaki, 2008). Descriptive statistics describe the size and distributions of

various elements in a sample. I evaluated variable distributions using frequencies, means

and standard deviations (J. Cohen, 1988; Tavakol & Dennick, 2011).

Tests for Association – Role and Clinical Decisions I tested for association between practitioner role, based on one of the three groups

that comprised the survey sample, and each of the four types of clinical decisions. I used

SPSS version 22 to complete a logistic regression to examine the relationships between

role and each group of clinical decisions; each group of clinical decisions served as my

dependent variable. I used a 95% confidence interval. I selected the logistic regression

because I created two category variables from the five category variable responses for the

clinical decisions. In the instrument, I asked respondents to indicate the frequency with

which psychosocial factors influenced diabetes clinical decisions. I used a five point

Likert scale (5-Always, 4-Often, 3-Sometimes, 2-Rarely, 1-Never).

To test for differences between the physicians and other survey respondents for

the four groups of clinical decisions, I defined the clinical decision as my dependent

variable. I grouped the nurse practitioners and the diabetes educators together since there

is overlap in the types of practitioners included in the samples. Since I compared two

groups, I used the t-test of difference between means (“independent samples t-test” in

SPSS). I excluded from my analysis any response that indicated “N/A – I don’t make or

influence these decisions”, an option for the frequency of influence on clinical decisions

questions. I considered the Likert scale to be an interval ratio variable, because the

difference from “Always” to “Often”, is the same as the “Rarely” to “Never”.

63

I also investigated perceptions of importance of psychosocial factors by testing

differences between each of the four groups of psychosocial factors and years of

experience, and physician specialty. In addition, I investigated difference in clinical

decisions influenced between each of the four groups of clinical decisions and years of

experience, and physician specialty. I used the one-way ANOVA to investigate

difference. I formed two groups for years of experience, 1) less than ten years, and 2)

greater than or equal to ten years. There were two physician specialties represented in the

primary care subsample of survey participants: family medicine and internal medicine.

Years of experience and physician specialty served as my independent variables.

Response Rates and Representativeness Response rate is the total number of surveys returned divided by the total number

of surveys sent. Calculating response rates for emailed surveys must also describe the

number of individuals who may have received it (Colbert, Diaz-Guzman, Myers, &

Arroliga, 2013). I kept close account of the entire sample. I used individual email

addresses for the primary care physicians and the nurse practitioners. I did not have

access to email addresses for the diabetes educators. Overall response rate was 29.8%, as

shown in Table 3.5.2. The response rate for the physicians was 39.8%, 16.0% for nurse

practitioners, and 35.8% for diabetes educators.

Table 3.5.2 – Survey Response Rates

Total Nurse

Practitioners Diabetes

Educators Primary Care

Physicians Total Sent 769 257 399 113 Total Responses 229 41 143 45 Response Rate 29.78% 15.95% 35.84% 39.82%

64

3.6. Risk Management There were several risks associated with this study related to recruiting subjects,

credibility and transferability, and content validity. Subject recruitment and participation

was critical to the success of the project. To mitigate the risk of inadequate data

collection, I communicated the aims of the study and the importance of study participants

to potential participants. I leveraged my professional network for interview recruitment,

which included the network of my dissertation committee members and colleagues at the

Ann Arbor Veterans Administration Center for Clinical Management Research. I also

used survey champions in each of the three survey sample groups to help maximize

response rates for the online survey.

Qualitative research is subject to threats to credibility and transferability. For this

project, creditably threats included omission and observer effects. Omission could occur

if I omitted clinical or patient situations relevant to explain how psychosocial factors

influence clinical care decisions. Observer effects could occur if participants indicate

what they think I want to hear, versus describing what actually takes place (Schensul &

LeCompte, 2012). I minimized these risks through my interview probes.

Validity Although validity has a common definition, there is no consistent method to

measure it. Validity refers to the extent to which the survey instrument accurately reflects

the intended construct (Chen & Paulraj, 2004; Friedman & Wyatt, 2006; Groves et al.,

2011). External validity is dependent upon the survey sampling method. The sample is

representative of the larger population (K. Kelley, Clark, Brown, & Sitzia, 2003).

Because I am interested in practitioners who make, or have input into, T2DM care

decisions, I ensured external validity by using a sample that draws from national and

65

regional associations, including practitioners from/with various: geographic areas,

practice settings, years of experience, and roles. Content validity refers to the extent to

which the survey questions reflect specific topics under investigation. Content validity

could be threatened if the survey instrument does not fully reflect the categories

identified in the interview data. I mitigated this risk by ensuring that the survey

instrument was informed by the interview data analysis, and by a lengthy cognitive

interviewing process. In addition, I ensured that the survey instrument reflected leading

practices found in the literature (see Appendix J: Principles of Survey Instrument Design)

(D. A. Cook & Beckman, 2006).

Institutional Review Board Approval The University of Michigan Health Sciences and Behavioral Sciences

Institutional Review Board (IRB) approved the study on February 7, 2014 (eResearch ID

HUM00085503; OHRP IRB Registration Number: IRB00000246).

66

CHAPTER 4

ACCESSING PSYCHOSOCIAL INFORMATION FOR CLINICAL

DECISIONS

If you look at people who seek a lot of care in American cities for multiple illnesses, it’s usually people with a number of overwhelming illnesses and a lot of social problems, like

housing instability, unemployment, lack of insurance… ― Paul Farmer

4.1. Introduction In this chapter, I describe how psychosocial information is accessed in the course

of providing diabetes care, specifically in making type 2 diabetes (T2DM) care decisions

in the outpatient setting. The study is comprised of two groups of participants: 1)

seventeen physicians who participated in one on one interviews, and 2) 219 online survey

participants sampled from a group of primary care physicians, nurse practitioners, and

diabetes educators. In this chapter, I address two of my six research questions:

RQ1: Which psychosocial factors do practitioners perceive to be important in making, or providing input into, care decisions for adult, type 2 diabetes patients? What is their relative priority?

RQ2: How do practitioners access psychosocial information?

The overall purpose of this investigation is to understand and document what

psychosocial factors practitioners consider in making, or providing input into, clinical

care decisions. I place psychosocial factors into four groups: 1) sociodemographic, 2)

psychological, 3) social relationship and living conditions, and 4) neighborhood and

67

community. Please see Appendix B: Psychosocial Factors for a further description of

these four groups of psychosocial factors. I group clinical decisions into four types: 1)

establishing appropriate levels of control, specifically for HbA1c (i.e., treatment goals),

2) prescribing medications, 3) making referrals to specialty care and support services, and

4) recommendations for diet, physical activity, and frequency of clinical visits. Please see

Appendix C: Type 2 Diabetes Clinical Decisions for a further description of these four

types of T2DM clinical decisions. Also in this chapter, I introduce my initial conceptual

model of psychosocial information access, which depicts 1) the top psychosocial factors

indicated by study participants, 2) their sources of psychosocial information, and 3) how

this psychosocial information is accessed.

4.2. Characteristics of Study Participants

Interview Participants A total of seventeen physicians were interviewed, as shown in Table 4.2.1.

Approximately half of the interview participants are family medicine physicians; the

other half are internal medicine physicians. There is one endocrinologist. There are

roughly equal proportions of male and female physicians; half have less than ten years of

experience. Approximately one quarter work in one of the following three practice

settings: 1) community clinics, 2) the Veterans Administration, and 3) public hospital

clinics.

68

Table 4.2.1 – Interview Participants Characteristics (n=17)

n

Specialty Family Medicine 8 (47.1%) Internal Medicine 8 (47.1%) Endocrinology 1 (5.9%)

TOTAL 17 Gender

Female 8 (47.1%) Male 9 (52.9%)

TOTAL 17 Years of Experience

<= 10 9 (52.9%) 11 – 20 6 (35.3%) 21 - 30 1 (5.9%) >= 31 1 (5.9%)

TOTAL 17 Practice Setting

Community Clinic 4 (23.5%) Federally Qualified Health Center (FQHC) 2 (11.8%) Veterans Administration 5 (29.4%) Public Hospital Clinic 4 (23.5%) University Hospital Clinic 2 (11.8%)

TOTAL 17

Online Survey Participants A total of 219 healthcare practitioners were eligible for and responded to the

online survey, as shown in Table 4.2.2. Approximately 20% primary care physicians, and

25% are from one of each of the following roles: 1) nurse practitioners, 2) registered

nurses, and 3) registered dietitians. Just over half of participants are certified diabetes

educators. Approximately 25% are family practice specialists, and 25% are internal

medicine specialists. Just over 15% have endocrinology as a specialty. The vast majority

practice in an outpatient clinic or a group practice.

69

Table 4.2.2 – Professional Characteristics of Survey Respondents (n=219)

Genesis Primary Care Physicians

North Texas Nurse

Practitioners

Michigan Diabetes Educators

TOTAL

Role Physician (MD) 39 - - 39 (17.8%) Physician (DO) 4 - - 4 (1.9%) Physician Assistant (PA) - - 2 2 (0.9%) Pharmacist - - 6 6 (2.8%) Nurse Practitioner (NP) - 39 12 51 (23.3%) Registered Nurse (RN) - - 58 58 (26.5%) Registered Dietician (RD) - - 58 58 (26.5%)

TOTAL 43 39 136 219

Certified Diabetes Educator

Yes 1 1 113 115 (53.2%) No 41 38 22 101(46.8%)

TOTAL 42 39 135 216

Clinical Specialty Family Practice 12 20 27 59 (27.4%) Internal Medicine 30 7 13 50 (23.3%) Endocrinology - 3 32 35 (16.3%) Podiatry - 1 - 1 (0.5%) Nephrology - - 1 1 (0.5%) Emergency Medicine - 3 - 3 (1.4%) Obstetrics/gynecology - - 1 1 (0.5%) Hospitalist - - 3 3 (1.4%) Other 0 5 57 62 (28.8%)

TOTAL 42 39 134 215

Practice Setting* Community Health Center 2 10 23 35 Free Clinic 7 6 14 37 FQHC 1 1 10 12 Outpatient Clinic 20 15 123 158 Veterans Administration 5 1 9 15 Home Health 5 9 18 32 Indian Health Service - - 9 9 Group Practice 45 18 44 107 Other 4 6 23 33

TOTAL 89 66 273 428 * select all that apply

70

The majority of participants are female; 20% are male, as shown in Table 4.2.3.

However the majority of the physicians are male. Just over half are between 45–64 years

old. Just over one third have ten years or less of experience, and another third have

between 11 and 20 years of experience.

Table 4.2.3 – Demographic Characteristics of Survey Respondents (n=219)

Genesis Primary Care Physicians

North Texas Nurse

Practitioners

Michigan Diabetes Educators

TOTAL

Gender Female 8 33 133 174 (80.9%) Male 34 6 1 41 (19.1%)

TOTAL 42 39 134 215

Age 25-34 1 6 7 14 (6.5%) 35-44 7 8 17 32 (14.9%) 45-54 19 14 28 61 (28.4%) 55-64 8 10 75 93 (26.2%) 65-74 6 1 7 14 (6.5%) 75-84 1 - - 1 (0.5%)

TOTAL 42 39 134 215

Years of Experience

<= 10 9 21 49 79 (38.7%) 11 – 20 15 10 43 68 (33.3%) 21 - 30 10 5 25 40 (19.6%) >= 31 7 - 10 17 (8.3%)

TOTAL 41 36 127 204

4.3. Overview of Findings The consideration and consistent influence of psychosocial factors emerged

immediately in the investigation, imparted in the initial interviews, and reiterated

throughout the subsequent interviews. Participants indicated that psychosocial factors are

consistently considered in making T2DM clinical care decisions. Since the physician

interview participants consistently emphasized the broad importance of psychosocial

71

information in general for providing patient-appropriate care, I describe them through

analysis of the online survey data, using the average Likert scores for each of the

psychosocial factors. In reporting my survey findings, I isolate physician survey

responses from the entire survey sample to support the goal of examining potential

differences between practitioner groups.

Summary of the Average Likert Scores for Psychosocial Factors Survey respondents indicated that the top psychosocial factor is financial strain

from the sociodemographic group, as shown in Table 4.4.2. Respondents indicated 4.84

out of a 5 point Likert scale (5 – Very Important, 4 – Important, 3 – Neither Important

nor Unimportant, 2 – Unimportant, 1 – Very Unimportant). Also, 99.4% of responses

indicated “Very Important” or “Important”. Next is mental health status (4.62/5; 97.6%),

and life stressors (4.57/5; 97.0%) from the sociodemographic group. Food security

(4.55/5; 94.8%) from the neighborhood / community group is next. Social support is next

(4.53/5; 97.5%), followed by health literacy (4.53/5; 97.0%), both from the psychological

group of psychosocial factors.

Summary of Sources of Psychosocial Information Among all respondents, 43.0% indicated that the patient is the most frequent

source of psychosocial information while among physician respondents, 49.5% indicated

as such, as shown in Table 4.5.2 (multiple responses were possible). The family and/or

caregivers is the next most common source, with 28.4% of all source selections among

all respondents, and 39.7% of all sources among physician respondents. Other providers

is the next most commonly chosen source, representing 15.5% of all source selections

among all respondents and 7.3% of all sources among physician respondents. The

72

electronic health record is the least common specifically-chosen source of psychosocial

information, reflecting 11.6% of source choices among all respondents and only 3.2%

these choices among physician respondents.

Initial Conceptual Model of Psychosocial Information Access To summarize the results for RQ1 and RQ2, I created an initial conceptual model

to depict the top 25% of psychosocial factors, practitioner role, the sources of

psychosocial information, and how the psychosocial information is accessed (see Figure

4.3: Initial Conceptual Model of Psychosocial Information Access). I build upon the

model in chapter six, where I show the final, complete conceptual model.

Top PFs Source of PI How Accessed

1. Financial Strain (4.84)*

3. Life Stressors(4.57)

5. Social Support(4.53)

2. Mental Health(4.62)

4. Food Security(4.55)

PI – Psychosocial InformationPFs – Psychosocial Factors* - Mean of responses to Likert scale responses: 5–Very Important, 4–Important, 3–Neither Important nor Unimportant, 2–Unimportant, 1–Very Unimportant** - % of Total Source Responses. Respondent could indicate more than one source.

Patient(43.0%)**

Family/Caregiver (28.4%)

Other Providers(15.5%)

EHR(11.6%)

• Data fields: mental health, payor status

• Prompting

• Engaging others• Listening• Questioning /

Clarifying

• Questioning (open-ended)

• Listening

• Asking/Calling (nurse, pharmacist)

6. Health Literacy(4.53)

Practitioner Role (Physicians vs. Other)

Figure 4.3: Initial Conceptual Model of Psychosocial Information Access

73

I now transition to more detailed results, using this model to structure the

presentation of results. I describe my findings from the physician interviews, followed by

my findings from the online survey.

4.4. Importance of Psychosocial Factors RQ1: Which psychosocial factors do practitioners perceive to be important in

making, or providing input into, care decisions for adult, type 2 diabetes patients? What is their relative priority?

Both interview and survey participants expressed beliefs that psychosocial factors

are important considerations in the course of making clinical decisions for diabetes care,

offering extensive and consistent narrative explanations of their perspectives regarding

their importance. As many interview participants highlighted, psychosocial factors are

vital because the care regimen requires extensive responsibility of the diabetes patient to

make daily decisions concerning self-care behavior—specifically regarding dietary

practices, physical activity, medication behavior, and self-monitoring. The bulk of these

decisions are made, and subsequent actions are taken, outside of the clinical

environment—away from the support and guidance of the care team. Accordingly, self-

care behaviors are greatly influenced by psychosocial factors characteristic of patients’

lives outside of clinical environments. Based on their understandings of this, practitioners

attempt to understand patients’ life circumstances to inform care decisions. A family

medicine physician with almost 20 years of experience, stated:

Any chronic disease that requires that the majority of the management is done by the patient at home on a daily basis is gonna be completely embedded in these psychosocial situations, and diabetes is the number one … You have to [consider psychosocial factors all the time]. (P16)

74

Survey participants also support the key theme that psychosocial factors are

considered in making, or providing input into, clinical decisions for diabetes care. For the

top six psychosocial factors, they indicated at least 4.53 out of a possible five, indicating

that they believed these factors were vital considerations.

4.4.1. Physician Interview Findings The physicians interviewed noted that if diabetes patients have the tools and

support to manage their lives, then they tend to experience good outcomes; however, if

they experience barriers that prevent them from managing their day to day

responsibilities, their ability to perform necessary self-care practices could be impaired

— thus resulting in worse health outcomes. Therefore, psychosocial factors are central

considerations as the physician attempts to assess to what degree a patient is able to

manage themselves, their environment, and their treatment. A medical director of a

federally qualified health center (FQHC) located in a federal refugee resettlement area in

the Northeast United States, offered that diabetes care must always consider psychosocial

factors because efficacy of care is essentially influenced by to what degree each

individual patient is able to navigate the various barriers to self-care they may confront:

About 5 years ago, we tried to figure out why [a] sub-population of patients are consistently with A1Cs greater than 9 … we looked at demographic information, age, country of origin, male versus female. The only persistent factor that we found out of our population of 700 plus diabetics … [was that they] have anxiety, depression, schizophrenia, and ultimately people that have a poor control of what’s called “managing their lives” … [they] have obviously a poor control with managing their diabetes. (P01, Family Medicine)

As the above quote suggests, psychosocial information is important because it can offer

context and perspectives to physicians concerning the issues their patients may be

confronting, which may present barriers to care. For example, P17 states, “I have a very

75

hard time giving a diagnosis of non-compliant, because we’re looking at the tip of the

iceberg, we don’t know what’s going on in their life … you’re gonna look at psychosocial

factors” (P17, Family Medicine).

Physicians highlight the difficulties that their patients have with managing

diabetes because of the demands placed upon them. Participants spoke to the

considerable strains of the all-encompassing care regimen. An internal medicine

physician (P02) with twenty years of experience practicing in urban areas, states,

“diabetes is probably the most difficult of diseases … [it requires] management all your

life. It affects every facet of your life”. The patient must undertake relatively strict, daily

self-care practices such as specific dietary choices—which can be difficult if a patient is

in an insecure or chaotic living situation. The importance of consistency with timing or

type of food is therefore one disease-related factor that makes diabetes more difficult to

manage, for the patient and for the physician:

If you have significant diabetes where you need to be on insulin … then you really need to be pretty regiment[ed]. You kind of eat the same amount of carbs every day, and every meal. Otherwise you can’t figure out, you can’t know how much insulin you should be taking ... It needs to be at the same time every day or things get out of whack pretty quickly. (P16, Family Medicine)

Next, I highlight participants’ insights into the importance of specific

psychosocial factors shown in the Conceptual Model of Psychosocial Information

Access, which physicians addressed in the interviews as important considerations in

making clinical decisions concerning their T2DM patients: 1) financial strain, 2) mental

health status, 3) life stressors, 4) food security, 5) social support, and 6) health literacy.

Then, I explain how each helps physicians understand potential barriers and facilitators to

self-care, in order to tailor the treatment they believe most appropriate for the patient.

76

Financial Strain There are a myriad of reasons why a patient may not be following their

recommended diabetes care regimen, and physicians believe financial barriers are

important to consider as a possible explanation. An internal medicine physician (P09)

with twenty-five years of experience based in Texas, states how important financial

barriers can be; “How much disposable money do people have is a really big thing, and

something I take into consideration … I don’t care if they speak English or they speak

Spanish or Ethiopian or whatever, a more common denominator is just poverty.”

Physicians explain that patients tend to prioritize their self-care behavior relative

to their basic needs, which can be threatened by financial strain. A family medicine

physician (P14) with extensive experience with at-risk patients, shared how patients may

focus their energies and attention on meeting such basic needs ahead of managing their

diabetes: “lights, water, cable, internet… phone … all those things … because if they

don’t have that, nothing else matters. They’re just trying to get the lights on, ‘Don’t even

talk to me about my diabetes, because I’m trying to get my lights on.’” An internal

medicine physician (P08) with 15 years of practice experience, further stresses, “if you’ve

got limited resources getting the kids to school and getting them the things that they need,

[that] takes priority over getting the healthy foods to manage the diabetes.”

Thus when basic needs are not being met, patients tend to shift their priorities

away from recommended self-care practices, and towards behavior that helps them meet

their basic needs. Financial strain can cause such a shift in priorities.

Furthermore, physicians noted that when basic needs are not being met, patients

may believe that their diabetes management is important, but that this belief may not be

77

easily transformed into self-care behavior. Again, financial strain can cause a shift in

priorities in a way that impedes self-care, as summarized by this physician:

Is it because they literally don’t have enough money to pay for their medicine? They might think it’s really important, but they’ve made the decision that, “I need to have food for my child versus getting my Januvia [diabetes tablet medication] which is too expensive. (P01, Family Medicine) Therefore, as physicians note, financial barriers can cause a patient to miss

appointments, not take medications, or make unhealthy dietary choices. P09 shared

phenomenon was evident among his/her patients, “[access to money] affect[s] whether

or not people can buy the medicines”. Or, as this Texas-based family medicine physician

(P10), describes: “they may not … have the funds … to buy healthier type foods and so

they’re eating on the run.”

An internal medicine physician (P08) based in Michigan states how financial

strain can present barriers to self-care, which then can result in unplanned hospital

admissions; “we recently had somebody with 2 hospital admissions in 1 month, because

she’s on 15 medications. She has a really big co-pay. She doesn’t get paid at the same

time her medicines are due. And so she’ll go a few days without taking her medicines.”

Physicians with experience treating certain patient populations, based on

geographic location or care setting, consider financial strain to be relevant for all of their

patients. For example, financial barriers are a consistent consideration for physicians who

see under- or uninsured (“self-pay”) patients. Physicians with such patient populations

simply assume that financial barriers are always a threat to a patient’s ability to perform

consistent, healthy self-care behavior. A family medicine physician (P10) with almost a

decade of practice experience in a low-income, community clinic, shared, “the financial

barrier … is pretty much universal here at our clinic.”

78

Mental Health Status Mental health issues are important because they can indirectly contribute to

barriers to healthy self-care; thus, similar to financial strain, mental health issues impede

self-care. An internal medicine physician (P07) states, “The reason … is not always just

socioeconomic. Often it’s mental health issues or something as well, that are preventing

them from being compliant.”

Mental health issues present barriers to self-care by restricting the patients’

motivation and requisite understanding of treatment recommendations. This can result in

unhealthy self-care behavior. A family medicine physician (P17) and medical director of

a federally qualified health center (FQHC) in the Midwest United States explains, “if

they’re depressed … they are not able to check their blood sugar. Of course if they don’t

check, guess what, they don’t know what it is … they’re not motivated to do any

exercise.” These barriers can be exacerbated by comorbid conditions, as expressed in this

patient case, “He’s got schizophrenia … He is cognitively impaired.… He smokes despite

his very end-stage COPD (chronic obstructive pulmonary disease) and diabetes” (P16,

Family Medicine).

Some physicians consider mental health status for all patients. The prevalence of

mental health issues cause them to be broadly considered:

Every doctor has to think about [mental health] … 15% of the U.S. population has depression, so everyone should think about depression, really … And SSRIs [Serotonin-Specific Reuptake Inhibitors, typically used to treat depression] were once the number one prescribed medication in the United States. (P06, Family Medicine)

Since the diabetes self-care regimen requires considerable responsibility of the patient,

mental health issues are consistently noted. An internal medicine physician (P07) with

over a decade of experience with at-risk patient populations makes the point this way,

79

“certainly if people have mental health conditions … that is clearly going to stand in the

way of their ability to care for themselves.”

As with financial barriers, mental health issues tend to be considered when

treating certain patient populations, based on geographic areas where physicians practice.

Physicians describe, based on their practice experiences, how patients from certain

geographic areas tend to have higher concentrations of mental health issues. One family

medicine physician with practice experience in different, major United States urban

centers states, “There’s a lot of mental illness in, I think urban kind of violent, or nearly

violent environments” (P06). Another family medicine physician (P16) describes how

economic challenges are common among her patients with mental health issues, therefore

they tend to gravitate to low-income communities; “Ypsilanti (Michigan) has a very large

population of schizophrenic and bipolar disease patients. It’s a cheap place to live…. so

people who don’t have very much money go live there. And people with schizophrenia

don’t have very much money.”

Life Stressors Life stressors are important considerations because they can have dramatic and

negative effects on self-care behavior. Physicians try to assess stress level to get an

understanding of what the patient will realistically be able to do. Awareness of life

stressors, and association with other psychosocial factors, also help physicians understand

potential reasons for unhealthy self-care practices. This understanding regarding drivers

of self-care is important in helping them determine the type of support a patient might

need.

80

Stressors are generally associated with other psychosocial factors, which can stem

from various everyday circumstances, often concerning difficult relationships with other

people such as family members. As this physician noted, he considers “stressors in terms

of do they have unstable relationships with their family members” (P01, Family

Medicine). Another family medicine physician (P06) with experience with at-risk

patients provides examples of stressors that impede self-care he has observed with his

diabetic patients, “stressors in their lives … a family member who's in jail …”

One physician also highlighted the problem of gun violence in the broader

community as an important source of stress for her diabetic patients. Physicians

recognize that stress affects self-care because it may lead to self-soothing and impaired

decision-making “stressors … causing them to overeat, make poor decisions” (P14,

Family Medicine). Life stressors, again accompanied by other psychosocial factors, also

tend to lower a patient’s motivation and confidence: “[lack of] support … financial

resources... general mental health… stressors. Certainly go hand in hand with

motivation, self-efficacy ... patient's sense of agency” (P15, Family Medicine). These

experiences of stress can also increase vulnerability to other problems, such as mental

illness, as one physician argued: “stress … loneliness … things which aggravate to

depression” (P11, Internal Medicine).

Some practice settings tend to have patients who experience various, persistent

life stressors — often associated with poverty. A family medicine physician who sees

patients at a community clinic shares, “In this setting, we have a skewed population here

… in our setting in general …. uninsured, poor people, you have to always consider that

… they have a very high stress level in their lives and that stress affects them negatively

81

… they … come in with depressive-type symptoms or stress-related symptoms” (P10,

Family Medicine).

Food Security Lack of consistent access to food is important because it can present barriers to

following dietary recommendations. Physicians described two threats to food security: 1)

not enough food due to financial strain, and 2) lack of access to healthy foods in

neighborhood. Financial strain is a very severe problem with many layers, including food

access, as articulated by a family medicine physician (P16):

Homeless patients have a real hard time managing their diabetes, things like … some consistent kind of food … you [can’t] fully assess the quantity of carbs that you are ingesting, and have a reasonable way of figuring out how much of insulin you should be taking. So for people who need insulin to manage their diabetes, being homeless makes it incredibly difficult.

A family medicine physician (P15) also states that food insecurity is an important issue

standing in the way of self-care among some of her patients, “… the reason why their

A1c is high is because they don't have enough food, or if they're buying a lot of calorie-

dense, nutrient-poor food.”

Consideration of this issue appears to be situational. The family medicine

physician continues, describing what prompts her consideration of food insecurity, and

how other psychosocial factors may also be considered, “if someone is uninsured or

Medicaid, I'm certainly gonna be … asking a lot more questions about financial security,

and food security ... and like, are needs getting met” (P15).

Social Support Following the diabetes care regimen requires considerable social support; it can

act as a facilitator or barrier for adherence to recommended self-care behaviors.

82

Understating the level of social support available to the patient helps the physician

understand to what degree the patient may have available to help them, “[understanding

level of] social support [is] definitely [important].… And … obviously, how many people

live in the family is also important. If you’re by yourself versus you have a family to

support” (P11, Internal Medicine). As will be discussed in the next chapter, this

assessment of available help is important in assessing both patient capabilities and in

evaluating the causes of patient behavior. The importance of social support to diabetes-

related self-care was summarized by this physician: “Diabetes is really difficult to

manage all by yourself … doing the shopping, doing the meal preparation, making trips

to the pharmacy … Oftentimes, it’s a team thing. The spouse or the significant other or

oftentimes, the child is the one going to the pharmacy or going to the store” (P01, Family

Medicine).

As suggested above, strong social support can facilitate self-care. An internal

medicine physician (P13) with practice experience with at-risk patients in the East and

Midwest United States, further illustrates with reference to a specific patient case:

And I know, for example, that he [diabetic patient] has a very supportive wife who will do whatever needs to be done at home. Whatever it is I say that he should do, she will make sure it happens, and I know that’s a key part of his psychosocial environment.

Conversely, physicians also share specifically how lack of social support can act as a

barrier to recommended self-care behavior. Low social support can impede the patient’s

ability to manage the various self-care responsibilities associated with the diabetes care

regimen. As exemplified in the following quote, physicians note when a patient’s social

relationships may present barriers to self-care, such as friends, neighbors or roommates:

83

I’ve got one gentleman … [he] always seems to have some problematic person renting a room from him … that is very disruptive to his ability to manage his chronic conditions because this person’s always causing trouble, it causes him a lot of stress. They’re bringing substances into the house, there’s always alcohol around. There’s a lot of drama. Those kind of things make it really difficult to manage [diabetes]. (P13, Internal Medicine) Social support is considered in specific situations. They consider potential barriers

to self-care due to low social support when they see patients who may not have family

members or friends who can help with self-care demands. An internal medicine physician

(P13) describes this when he considers the patient’s support network; “I think people’s

social support network is really important. So, do they have caregivers, friends, other

people in their social network that can support them? Or conversely, that make it hard

for them to effectively manage their diabetes?” Understanding a patient’s level of social

support is critical to approximating what they can realistically perform.

Health Literacy Health literacy is an important consideration. Physicians assert that when diabetes

patients understand the treatment, they are more prone to follow recommended self-care.

Consequently, they experience better outcomes. As this family physician said: “… some

people just don't understand the health problem … they could have low health literacy …

they [may] not focus on that aspect of their health” (P06, Family Medicine).

An internal medicine physician (P13) discusses how a patient’s level of health

literacy and health numeracy might affect their ability to understand the rationale

underlying treatment decisions: is important:

Their level of health literacy, their level of health numeracy. Those are things that I think a lot about… it relates to their ability to understand the numerical things about diabetes and … risk for things in the future and how that may relate to their understanding of sort of time perspective and reducing the risk of things over a five-day or 10-year time horizon.

84

Limited health literacy can also affect patient motivation. An internal medicine

physician (P13) asserts this connection as follows: “their health literacy or their

numeracy … obviously, there's some sort of correlation... a fairly strong correlation there

… we … use education as a proxy.... a higher level of education, in many cases [is]

correlated with things like self-efficacy, activation, treatment engagement, motivation.”

Physicians consider health literacy issues when caring for at-risk patients. A

family medicine physician (P11) describes her experience with at-risk patients, “a lot of

[our patient] population … is illiterate. But if you give them health education, they

wanna learn and take care of themselves … health education probably is one most

important factor [in these patients doing well].” An internal medicine physician

considers health literacy issues when he sees patients who may have low levels of formal

education, which, in his experience, influence how involved they may be in their care,

“people with lower levels of education are more often people who have lower levels of

self-efficacy, motivation, and treatment engagement” (P13, Internal Medicine).

4.4.2. Survey Findings Since there are no previous published studies nor an established standard

procedure to select the top psychosocial factors, I focus on the top 25% (6 of 23)

reflected in my sample. To identify the top 25% of psychosocial factors, I list them

according to highest average score on the Likert scale for my sample (5 – Very

Important, 4 – Important, 3 – Neither Important nor Unimportant, 2 – Unimportant, 1 –

Very Unimportant). Please see Appendix K: Average Likert Scores of Psychosocial

Factors for the complete list of 23 psychosocial factors I investigated, sorted by highest

to lowest average score on the Likert scale.

85

The psychosocial factor, financial strain, had the highest average Likert score of

all psychosocial factors measured. As shown in Table 4.4.2 – Average Scores of

Individual Psychosocial Factors, the six highest average scores for psychosocial factors

are, in order of relative priority: 1) financial strain (sociodemographic), 2) mental health

status (psychological), 3) life stressors (psychological), 4) food security (neighborhood /

community), 5) social support (social relationships / living conditions), and 6) health

literacy (psychological). I isolate physician responses in the table. Please see Appendix L

Average Likert Scores of Psychosocial Factors by Group for a listing of average scores of

the psychosocial factors by the four groups: 1) psychological, 2) social relationships /

living conditions), 3) neighborhood / community, and 4) sociodemographic.

The differences for all six of the top psychosocial factors indicated are statistically

significant based on the comparison of physicians to other practitioners. I completed an

independent-samples t test comparing the means of the Likert scores of importance (5 –

Very Important, 4 – Important, 3 – Neither Important nor Unimportant, 2 – Unimportant,

1 – Very Unimportant) between the two groups: 1) physicians, and 2) the nurse

practitioners and diabetes educators.

For financial strain, I found a significant difference between the two groups

(t(162) = 2.657, p < .05). For mental health status, I found a significant difference

between the two groups (t(162) = 4.531, p < .05). For life stressors, I found a significant

difference between the two groups (t(162) = 3.978, p < .05). For food security, I found a

significant difference between the two groups (t(157) = 8.354, p < .05). For social

support, I found a significant difference between the two groups (t(157) = 4.029, p < .05).

Last, for health literacy, I found a significant difference between the two groups (t(162) =

86

-4.897, p < .05). I report the absolute value of the t-value. In all cases, the means for the

nurse practitioners and diabetes educators are higher than the physicians for each

psychosocial factor — suggesting that, on average, these practitioners viewed

psychosocial factors as more important than their physician counterparts.

Table 4.4.2 – Average Likert Scores of Individual Psychosocial Factors

Total (n = 164)

Primary Care Physicians

(n = 36)

Nurse Practitioners & Diabetes Educators (n = 128) p value

Financial Strain 4.84 (383) 4.69 (.525) 4.88 (.323) .009 Mental Health Status 4.62 (.556) 4.28 (.701) 4.73 (.465) .001 Life Stressors 4.57 (.576) 4.25 (.649) 4.66 (.522) .000 Food Security 4.55 (.633)a 3.88 (.729)b 4.74 (.460)c .000 Social Support 4.53 (.572)a 4.20 (.678)d 4.62 (.504)e .000 Health Literacy 4.53 (.580) 4.14 (.683) 4.64 (.498) .000

Note. Respondents were asked to indicate the importance of psychosocial factors in making, or providing input into, diabetes care clinical decisions. Responses were captured in a Likert scale: 5 – Very Important, 4 – Important, 3 – Neither Important nor Unimportant, 2 – Unimportant, 1 – Very Unimportant. Standard deviations listed in parentheses. a n= 159. b n = 34. c n = 125. d n = 35. e n = 124.

4.4.3. Comparison of Results Both interview and survey participants indicated that psychosocial factors were

influencers of clinical care decisions. Five of the top six psychosocial factors, as

indicated by average Likert score in the survey, were also discussed extensively by the

interview participants: 1) financial stress, 2) social support, 3) mental health status, 4)

health literacy, and 5) life stressors. However, neighborhood context was emphasized

more by interview participants than survey respondents; while interview participants

mentioned this as an important theme in the context of life stressors, it was one of the

lower-scored factors in the survey (4.01/5: 76.1%). Additionally, although survey

87

participants emphasized food insecurity as a psychosocial factor considered (4.55/5:

95.0%), it was not as commonly mentioned by interview participants – although it was

mentioned enough to include it as a theme. Differences between practitioner roles will be

discussed further in the discussion section.

4.5. Sources of Psychosocial Information RQ2: How do practitioners access psychosocial information? Both interview and survey participants indicate that the patient is a significant

source of psychosocial information. Interview participants conveyed how the patient, via

the patient consultation, is the primary source of psychosocial information. Both sets of

participants also shared that the patient’s family and caregivers are vital sources of this

information. Please see Table 4.5.1 for a summary of the sources of psychosocial

information, and detail on techniques physicians use to access it.

Table 4.5.1 –Sources of Psychosocial Information, How Accessed

Source of Information How Accessed • Patient • Questioning (open-ended)

• Listening

• Family/caregivers • Engaging others • Listening • Questioning / Clarifying

• Other Providers • Asking / calling them (i.e., nurse, pharmacist - members of care team)

• EHR • Data fields: mental health, payor status • Prompting

Participants also use the electronic health record as a source for psychosocial

information, although less commonly per the survey results. Interview findings showed

88

that the EHR was most helpful for identifying a small number of factors of interest to

them: mental health status, medication refill behavior, and attendance at clinical

appointments (see Figure 4.3). More detail is provided below.

4.5.1. Physician Interview Findings

Patients Interview participants indicated that patients are their main source of psychosocial

information; it is primarily acquired when physicians ask patients about barriers to

following the diabetes care recommendations that they may be experiencing (see

Questioning (open-ended) on Figure 4.3). As this participant explained, “I will ask them

point blank, how is their mood and how they’re doing … I usually will just rely on the

patient telling me” (P07, Internal Medicine). Beyond eliciting general issues, some

physicians also have direct conversations with their patients about specific psychosocial

issues, if they believe they may be experiencing barriers because of them. As this

participant explained, “I very much have conversations about what they’re able to afford

in terms of their medications” (P15, Family Medicine).

In addition to context on barriers, these conversations may also elicit specific

patient priorities, a key input in selecting clinical goals appropriate for the particular

patient’s circumstances. A family medicine physician (P16) recounts a recent

conversation she had with her patient concerning how employment demands interfered

with his ability to follow the diabetes regimen. As a result, his priorities shifted:

He said, “Look, my decision is to keep my job. I’d rather keep my job than manage my diabetes. I understand that the long term health consequences are bad, but I’m not gonna treat my diabetes right now. I’m just gonna let it go. And when I get on my feet again and I get a place to live again, then I’ll talk to you about my diabetes management.”

89

Some patients will also share psychosocial information about barriers they are

experiencing, without prompts from the physician, “So many times they just come tell

you, ‘I can’t afford that insulin’” (P03, Internal Medicine). In these cases, the primary

mode of access is Listening (see Figure 4.3). Of interest are issues they may be

experiencing in that particular moment, or those on a more long-term basis.

Caregivers and Family Members Caregivers and family members are also sources of psychosocial information for

physicians. Participants indicate that engaging others, especially the patient’s support

network is important (see Figure 4.3) since it can be a key information source regarding

psychosocial issues facing a patient. As this physician explained: “[for some patients] …

we have to engage the family. It’s really important to get some family member [to attend

clinic visit with the patient], or at least some decision makers in the family who play an

important role … that really helps” (P11, Internal Medicine).

Caregivers provide psychosocial information directly regarding factors about

which they have knowledge, like access to support; “[the] caregiver might tell you

[about] their home situation” (P03, Internal Medicine). They also can provide insight on

specific cognitive issues, which can be difficult for the physician to access from the

patient themselves; “So if they’re having problems remembering, they’re having memory

issues that the family tells us, ‘Oh, we’re noticing some things at home.’ You’ll get that

information from their family member caregiver. We have to rely on that” (P07, Internal

Medicine). This information may be accessed by physicians either by listening or by

questioning/clarifying (see Figure 4.3).

90

Other Practitioners Physicians rely on other members of the care team to help identify potential

psychosocial barriers, and help address them – primarily by asking/calling these

practitioners (see Figure 4.3). For example, physicians acknowledge that nurses can be an

excellent source of information on psychosocial barriers, and one described how he asks

them directly about his patients’ potential barriers, “nurses are totally onto this

[psychosocial barriers]. They're the ones who can answer a lot more [questions about

psychosocial barriers] than the providers” (P02, Internal Medicine). A family medicine

physician (P12) also shares how he calls the pharmacist when he might have concerns

about psychosocial barriers to medication behavior. He consults with this member of the

care team to confirm the barrier and develop approaches to address it, “[When] I have

concerns about adherence … which we would frequently do. We would call our

pharmacy and … .get [the] history … they're supposed to get refills every month. Well

she's had three refills in the last six months. Well that explains it…”

Electronic Health Record Although they are the most frequently relied upon information sources, patients

and caregivers may not fully articulate the psychosocial issues they are facing.

Accordingly, physicians may use the electronic health record (EHR) to help prompt a

conversation about psychosocial barriers (see Figure 4.3). Additionally, physicians

review the data fields to help understand medications refill behavior, attendance at

follow-up appointments, and explore other potential barriers (see Figure 4.3). In these

cases, they will use the EHR to help identify a specific psychosocial barrier, such as

mental health status.

91

EHR – Prompt for Further Probing An internal medicine physician (P07) states how he uses the EHR to check

specifically for medication behavior, “if they’re not refilling … you’ll know … there’s an

issue with meds. You’ll talk to them about it.” P07 also explains how insight regarding

medication refill history found in the EHR can trigger dialog probing for potential

psychosocial barriers; “you … look at their medication history … [to confirm] if they’re

not refilling … You’ll know … there’s an issue with meds. You’ll talk to them about it and

depending on what they say … [confirm if] they’re agreeable to seeing someone to help

them out… They might say, ‘I can’t afford it.’ Or they might say, ‘I forget,’ or ‘I don’t

need it,’ or ‘I don’t think I need it, I feel fine.’”

P07 also shares how information in the EHR helps her understand if the patient is

missing follow-up appointments, and this information helps guide the conversation with

the patient which could reveal psychosocial barriers, “information [in the record] …

could [prompt me to probe], ‘Why can’t [you] make it to the appointment? Why did you

miss your last three appointments?’”

EHR – Reviewing Data Fields The EHR is also used to help identify a problem. It can provide insights on

psychosocial factors that patients may not share directly. For example, they use the EHR

to help them assess mental health status; “generally people wouldn’t tell me about

suicidal ideation, but sometimes it would be noted [in the EHR] that they had a suicide

attempt” (P06, Family Medicine).

Payor status, contained in the clinical record, can also help guide the conversation

on potential self-care barriers due to financial strain, “if someone is uninsured or [has,]

Medicaid, I’m certainly gonna be, right away asking a lot more questions about financial

92

security, and food security, and housing status, and just kind of checking in and seeing

like, safety … I mean, just in terms of like housing security [if] needs [are] getting met”

(P15, Family Medicine).

4.5.2. Survey Findings In line with the interview findings, survey respondents indicate that the patient is

the most frequent source of psychosocial information–representing 43% of all source

selections among all respondents, and 50% among physician survey respondents, as

shown in Table 4.5.2. The table shows the absolute number of all sources indicated, for

each of the 23 psychosocial factors investigated. Respondents could indicate more than

one source for each psychosocial factor. The table also shows the percentage of all

selections for each source given as an option. The family and/or caregiver(s) is the

second most indicated source (28.4%), followed by other providers or members of the

care team (15.5%). The electronic health record is the least indicated specific source

(11.6%). The importance of these factors are also represented in Figure 4.3.

Table 4.5.2 – Selected Sources of Psychosocial Information

Total Selections (n = 164

respondents)

Primary Care Physician Selections

(n = 36 respondents)

Nurse Practitioners & Diabetes Educators (n = 128

respondents) Patient 1925 (43.0%) 340 (49.5%) 1585 (41.8%) Family / Caregivers 1270 (28.4%) 273 (39.7%) 997 (26.3%) Other Providers / Members of Care Team 696 (15.5%) 50 (7.3%) 646 (17.0%)

EHR 521 (11.6%) 22 (3.2%) 499 (13.2%) Other 37 (0.8%) 0 (0.0%) 37 (1.0%) No Reliable Source 28 (0.6%) 2 (0.3%) 26 (0.7%)

TOTAL 4477 687 3790 Note. Respondents could indicate that they relied on more than one source (“check all that apply”). Totals represent the total number of times the source was indicated, for each of the 23 PFs investigated.

93

4.6. Practitioner Characteristics – Physician Interview Findings

RQ5: What practitioner characteristics (i.e. role, age) are associated with their use of psychosocial information?

In this section, I address the fifth research question. I describe practitioner

characteristics that are associated with psychosocial information use. Diversity in amount

of experience and representation of three specialties in the physician interview sample

offers varied perspectives on the relative importance of psychosocial factors. Experienced

physicians tend to emphasize the necessity to consider the influence of psychosocial

factors; this is a fundamental part of how they deliver care. Less experienced physicians

understand their importance and agree on their relative priority. Family medicine

physicians express that those that chose the specialty may be more inclined to consider

psychosocial factors.

Experience Experienced physicians detail how their insights on the importance and influence

of psychosocial factors have developed in the course of providing patient care over

several years, which include innumerable patient cases. A family medicine physician

(P01) emphasizes that appreciation for the influence of psychosocial factors is essential

to sustaining a practice in certain care settings; “as a family practice doc working in a

community health center for a decade and a half, either you acknowledge the impact of

the psychosocial problems or you’re not gonna last.”

Physicians’ experience in providing care for diabetic patients informs the

importance they place on evaluating psychosocial information, which may have been

94

provided by other members of the care team. An internal medicine physician (P09) with

over two decades of experience states, “I think the difference between being a young

doctor and an old doctor, is that the young doctor will just take the history that somebody

else has gathered as gospel. An older doctor will look at it and go, ‘I wonder if all of

that’s true?’ Because they’ve seen enough patients to realize that it is not true.”

Family Medicine Specialty A family medicine physician asserts that the specialty itself may predispose

physicians to be sensitive to the importance of psychosocial factors:

I would hope that [residents have] a little bit of bias as a family practice resident that they care about the big picture and not just the individual disease. Because there’s inter-relationships between … co-morbid problems … and … their psychosocial issues. I mean it’s kind of … I don’t know, how does a plumber learn to be a plumber? I mean he watches a good plumber after a while, and then he goes out on his own, and goes back to that good plumber every now and then. They’re observing how I’m interacting with patients and trying to glean from it some sort of knowledge that they can bring to their own practice … we take care of a lot of diabetics … this is the way we do it. Here is our data to say whether or not it’s been successful or not. (P01, Family Medicine) In some cases, appreciation for the importance of psychosocial factors may have

started prior to medical training, depending on an individual’s experiences before

attending medical school. Those who believe these issues are important may have formed

these perspectives prior to beginning formal medical training. As this family medicine

physician (P15) shares, “I had [perspectives on psychosocial factors] going in just kind of

through personal experience.… I definitely kind of didn’t have a traditional path in

medical school … I’ve had a lot of other experiences and done international work and

did a lot of other things … done community organizing … [so] in medical school, for

sure, I was very much aware of it and very much aware of the way in which it was

impacting my patients” (P15, Family Medicine).

95

4.6.1. Survey Results

Likert Scores for Psychosocial Factors and Clinical Role As table 4.4.2 shows, there is a statistically significant difference between average

Likert scores for psychosocial factors between physicians on the one hand, and the nurse

practitioners, registered nurses, and diabetes educators on the other. The nurse

practitioners, registered nurses, and diabetes educators scored each of the psychosocial

factors higher than their physician counterparts.

Likert Scores for Psychosocial Factors and Years of Experience Based on interview participant perspectives, I further investigated perceptions of

psychosocial factors by testing differences between each of the four groups of factors and

years of experience. Years of experience is my independent variable and each of the four

groups of psychosocial factors is my dependent variable. I measured the independent

variable based on if the respondent had less than ten years of experience, or greater than

or equal to ten years of experience. I used the one-way ANOVA to investigate difference.

I found no statistically significant difference between years of experience and average

Likert scores for psychosocial factors for any of the four groups of psychosocial factors

(Psychological: F(1,162) = 0.12, p > .05; Social Relationships / Living Conditions:

F(1,157) = .005, p > .05; Neighborhood / Community: F(1,157) = 1.985, p > .05;

Sociodemographic: F(1,162) = .020, p > .05). In fact, I found remarkable similarities in

average Likert scores between the less experienced and more experienced practitioners.

Likert Scores for Psychosocial Factors and Physician Specialty (Internal Medicine and

Family Medicine) Also based on the interview results, I investigated whether there are differences

between average Likert scores for each of the four groups of psychosocial factors and

96

physician specialty. Specialty (Family vs. Internal medicine specialty) is my independent

variable and each of the four groups of psychosocial factors is my dependent variable.

The primary care physicians were comprised of two specialties: family medicine and

internal medicine. I used the one-way ANOVA to investigate differences only in the

physician sample (n=36). I found no statistically significant difference between specialty

and average Likert scores for psychosocial factors for any of the four groups

(Psychological: F(1,34) = 2.19, p > .05; Social Relationships / Living Conditions:

F(1,33) = 1.42, p > .05; Neighborhood / Community: F(1,32) = .005, p > .05;

Sociodemographic: F(1,34) = 5.17, p > .05).

These findings are reflected in the initial conceptual model of psychosocial

information access (see Figure 4.3) where practitioner role is depicted as a moderator of

perceived influence of information sources.

4.7. Conclusion In this chapter, I described the psychosocial factors that practitioners perceive to

be important in the course of making type 2 diabetes care clinical decisions in the

outpatient setting, which answers RQ1. As depicted in the initial conceptual model of

psychosocial information access (see Figure 4.3), the top six psychosocial factors are: 1)

financial strain, 2) mental health, 3) life stressors, 4) food security, 5) social support, and

6) health literacy.

I detailed how psychosocial information is accessed from four main sources: 1)

patient, 2) family / caregivers, 3) other providers, and 4) EHR. Practitioners indicated that

the patient is the most frequent source of psychosocial information, specifically via the

patient consultation. The patient’s family and caregiver(s) is also a source of

97

psychosocial information. These individuals offer critical insights on self-care activities,

such as medication behavior. The EHR is also a source, although less often.

I found a statistically significant difference in perspective on the importance of

psychosocial factors based on role. As a group, the non-physician roles (i.e., nurse

practitioner, registered nurse, diabetes educator) perceived psychosocial factors as more

important than their physician counterparts (see table 4.4.2). I found little difference in

perceived importance of psychosocial factors based on years of experience. I found no

statistically significant difference in perception of importance of psychosocial factors

based on physician specialty. Practitioner role is depicted on the initial conceptual model

depicted in Figure 4.3.

In the next chapter, I describe how psychosocial information is used to make

diabetes clinical decisions. I introduce a cognitive map depicting how practitioners make

clinical decisions and detail the specific clinical decisions for which psychosocial

information is used.

98

CHAPTER 5

USING PSYCHOSOCIAL INFORMATION FOR CLINICAL

DECISIONS

It is much more important to know what sort of a patient has a disease than what sort of a disease a patient has.

― William Osler (1849 – 1919)

5.1.Introduction In this chapter, I describe how psychosocial information is used in the course of

making type 2 diabetes (T2DM) clinical care decisions in the outpatient setting. I group

clinical decisions into four types: 1) establishing appropriate levels of control,

specifically for HbA1c (i.e., treatment goals), 2) prescribing medications, 3) making

referrals to specialty care and support services, and 4) recommendations for diet, physical

activity, and frequency of clinical visits. Please see Appendix C: Type 2 Diabetes

Clinical Decisions for a further description of these four types of T2DM clinical

decisions.

As described in Chapter four, both interview and survey participants indicate that

they consider specific psychosocial factors to be important in the context of diabetes care,

and that they use questioning, listening, asking and reviewing approaches to retrieving

information from their primary sources of psychosocial information: patients,

family/caregivers, other providers, and the EHR. To illustrate how practitioners use

psychosocial information, I introduce a cognitive map that summarizes their thought

99

processes when using this information. I also describe when and how psychosocial factors

influence clinical care decisions. I conclude this chapter with survey findings detailing

the influence of psychosocial factors on specific clinical decisions.

In this chapter, I address two of my six research questions:

RQ3: How do practitioners use psychosocial information? How does this information influence their specific care decisions?

RQ4: In which situations are psychosocial factors considered?

Summary of Findings

Interview Themes Physicians use psychosocial information to help them determine barriers to self-

care that patients may be experiencing. They use this information to inform clinical

decisions to address these barriers as appropriate for the patient’s circumstances.

Summary of Decisions Influenced by Psychosocial Information Psychosocial information is used most frequently for target level of control

decisions, as shown in Table 5.5.5.1. Respondents indicated 4.26 out of a 5 point Likert

scale ( 5 – Always, 4 – Often, 3 – Sometimes, 2 – Rarely, 1 – Never) in response to target

level of control decisions. Also, 81.9% of responses indicated “Always” or “Often.”

Making recommendations are next (4.18/5; 90.6%), followed by other decisions, (4.14;

80.8%), making referrals (4.13/5; 75.8%), and medications decisions (4.09/5; 88.6%).

The nurse practitioners, registered nurses, and diabetes educators indicated their decisions

were more frequently influenced by psychosocial factors than did the physicians. The

difference was statistically significant for medications decisions.

100

5.2. Cognitive Map of Psychosocial Information Use Cognitive maps provide a cognitive representation of concepts associated with a

particular area, showing the connections between them (Miles et al., 2014). I use the

cognitive map to answer my third research question.

RQ3: How do practitioners use psychosocial information? How does this information influence their specific care decisions?

I used this schema to depict how psychosocial information is used to make

diabetes care clinical decisions (see Figure 5.2). This visual representation helps

articulate “what is going through” physicians’ minds when they reflected upon their

experiences using psychosocial information to make diabetes care clinical decisions. I use

descriptive labels, drawn verbatim from the physicians’ interviews. In the cognitive map

of psychosocial information use, I use quotations and bold font to make physicians’

words conspicuous. For further detail on how I constructed the cognitive map, please

refer to chapter three, section 3.5.1.

101

Figure 5.2: Cognitive Map of Psychosocial Information Use

I now detail each of the five areas of the map, which represent the key concepts

that comprise the cognitive process of psychosocial information use: 1) considering

clinical practice guidelines (CPGs) in the context of the patient situation, 2) building and

maintaining rapport with patient, 3) triggers to gathering and using psychosocial

information, 4) assessing the patient, and 5) making the clinical decision.

5.3. Considering Clinical Practice Guidelines (CPGs) in Context of Patient Situation Physicians are aware of general clinical practice guidelines and consider them in

the course of clinical decision making, particularly for determining the appropriate target

for a patient’s HbA1c level. But they emphasize that they calibrate their use of guidelines

against several factors, including psychosocial ones. While they certainly consider the

102

guidelines, they assert that they are by no means a replacement for their own clinical

judgment for what they believe is best for the patient. Determining what decision is

“best” can be nuanced and imprecise, however. One internal medicine physician (P02)

attempts to quantify the ambiguity rooted in the various clinical decisions associated with

outpatient clinical care; “… I tell all the [internal medicine] residents ... ‘if you’re that

type of person who wants 100% clarity, you better go into surgery. Because in the

medical field, 70% is a good day. 70% certainty is a good day.’”

Evidence Physician participants stated that they consistently use the guidelines as a starting

point, as they are considered as general standards of care, and represent the best evidence

for appropriate targets. A family medicine physician (P01) with administrative

responsibilities for clinical performance metrics at a federally qualified health center

(FQHC) says, “I’m a big evidence-based medicine guy, so if there is good data behind a

clinical practice guideline, then either myself as an individual [and] as the Chief Medical

Officer of the clinic, I’ve tried to get my other docs to endorse those. So, I would say

we’re very heavily invested in evidence-based medicine.” Moreover, there are diabetes

guidelines considered beyond the HbA1c target recommendation—like target cholesterol

level, blood pressure, diet, physical activity, and an annual eye exam. These supplemental

guidelines appear to be considered holistically at each visit. A family medicine physician

(P10) described her consistent use of certain guidelines; “I feel like I always operate

considering guidelines because … there are certain standards that are just standard

medical practice … their LDL goal, are they taking Aspirin? Is their blood pressure well

controlled? I mean those are all clinical guidelines we look at that every single time we

103

see a patient. Have they received their Pneumovax vaccine? Have they received their flu

shot? Are they up to date on their other preventive measures?”

Not clear-cut Nevertheless, physicians expressed considerable ambivalence concerning the

guidelines. Although their clinical decisions are motivated by the evidence, their general

attitude is quite mixed regarding their use. This is especially so for diabetes guidelines

when compared to other chronic conditions. As one internal medicine resident (P04)

explains, “diabetes guidelines … they’re not really as clear-cut as [guidelines for]

hypertension … If … blood pressure is above a certain threshold, you need to do

something.…“

“They’re just guidelines”: HbA1c Target According to interview participants, establishing an appropriate HbA1c target is

the specific diabetes care clinical practice guideline that tends to most frequently

necessitate clinical judgment, influenced by psychosocial factors. As such, there are

“always exceptions” (P03). An internal medicine resident (P04) explains, “… we’re

always on the fence on changing things … we don’t push it that much.… Because there’s

a huge range of 7 and 8 … still there is no real [specific goal] … it’s more like there’s

clinical judgment rather than [a strict number] ... it’s not by numbers.” An internal

medicine physician (P03) explains how the guidelines are not directives; they are simply

aids to support his clinical decisions, “they’re just guidelines.” This sentiment is echoed

by an endocrinology fellow (P05), who contends that judgment is needed when

considering the guidelines for target HbA1c; “I think in the guidelines, there’s room for

clinical judgment, like a different A1c for people with hypoglycemia.” This requirement

104

for judgment in setting appropriate HbA1c goals is also expressed by more experienced

physicians. An internal medicine physician (P09) with over twenty years of experience

states, “I think there’s evidence to show that … moving someone below 9, does a better

job than leaving them above 9, even though you would not have achieved a goal of below

7.”

A1c trend more important than an absolute number For most physician participants, understanding a patient’s HbA1c over a period of

time is more important than the absolute number. Physicians stress the value of

understanding the patient’s HbA1c history in order to ascertain a trend. Knowing if the

HbA1c is decreasing, increasing, or remaining the same, is important; “Typically [if]

they’re trending down … like if the A1c has been 9, now it’s 8.5, or like 10 and now it’s

9, even if it’s not controlled. I mean we don’t do anything from there because it’s going

down.... if … the trend is in the right direction ... Then we’re fine with that, [I tell the

patient] ‘Keep doing what you’re doing’” (P04, Internal Medicine). However, even with

historical HbA1c data, the guidelines are unclear. A family medicine physician (P16)

described just how imprecise the decision to establish an appropriate HbA1c goal even

with trend data in hand can be:

It’s like “Okay, should I add insulin now or not?”… their A1c is ranging from say 8.6 to 9.5, not great control … by anybody’s ideas. The biggest decision is, do I start insulin or not … and it’s not clear-cut. It’s not like this patient’s A1c is 8.4 today, it was 8.4 last month, it was 8.4 the month before, it was 8.4 the month before that. I’d like it to be under 8 so I’m gonna add insulin this much and it’s gonna go to 7.9 and we’re gonna be done. People don’t work that way.… Their A1c goes all over the place and you’re sort of like looking for where’s the mean here and how much can I push down the top without going too low on the bottom. It’s complicated.

105

They Don’t Consider All Patient Circumstances Diabetes care decisions are based on various, dynamic factors — including

psychosocial factors — which physicians believe are not necessarily fully incorporated

into the guidelines. Their use of the guidelines hinges upon what information physicians

gather regarding the patient’s particular situation, primarily at the time of the visit.

Patient situations can be deeply rooted in psychosocial factors; “when you read through

all these guidelines, [they] probably cover most of the population. They try to cover

majority of it … it might be little bit difficult to apply all the guidelines, [depending upon]

patient population … their education level … [because] things are different [for certain

patients]” (P11, Internal Medicine). Physicians’ level of adherence to the guidelines is

therefore dependent upon the patient’s situation, derived from their understanding of how

a patient’s psychosocial factors may impact what they believe is an appropriate for them.

The most important factors are detailed below.

5.4. Build & Maintain Rapport with Patient As described in chapter four (section 4.5), the patient is the most frequent source

of psychosocial information for physicians; moreover, as was depicted in the initial

conceptual model of psychosocial information access (see Figure 4.3), this information is

gleaned almost exclusively as a result of the communication (open-ended questioning and

listening) that takes place in the patient-doctor relationship. Using this information is

dependent upon the physician’s ability to access it, which physicians express is heavily

dependent upon the level of trust in patient-doctor relationship. Trust enables the

physician to facilitate communication and commitment from the patient on various

diabetes care recommendations; “with any type of patient, you have to build a rapport

and you have to build that communication, and so that they buy in to what you’re

106

teaching them or what you’re recommending them, and that’s just basic patient-doctor

rapport and relationship” (P10, Family Medicine). Consequently, as depicted in Figure

5.2, physicians actively consider how to foster the necessary rapport and relationship.

Trust Physicians spoke extensively about the importance of building trust with their

patients, as this is what facilitates their access to, and subsequent use of, psychosocial

information. Patients may share quite intimate and sensitive psychosocial information

about their living situation and self-care capabilities. A family medicine physician (P15)

believes trust is integral to building the relationships with patients, “I think trust plays

into things a lot … If we have an established relationship, so they … know … that I care

about them.” Another family medicine physician (P14) explains how developing trust is

essential to how she practices medicine:

The way that I practice medicine … I know that I’m doing a lot more than just collecting information. I’m developing a relationship with them, I’m developing trust. I’m trusting them and they’re trusting me … it’s like everything comes together in the right way for this relationship to work, and then all of a sudden it’s like they understand that I actually really care about them. Because I actually care and I want them to be better. However, this information is rarely disclosed during initial visits, before the

relationship is established. An internal medicine physician (P09) with over a decade of

clinical experience describes establishing a trusting relationship this way, “It’s kind of

weird … unlocking of someone’s trust … you have to have trust in order to get to the

truth about someone’s psychosocial factors.”

Although physicians state that they attempt to start building trust at the very first

visit, they also recognize that the relationship must be built before they expect

transparency in disclosure of psychosocial information. The relationship develops over

107

several visits; “In that first visit, I think you have to be somewhat skeptical, because

people will usually try to paint a better picture … because they’re trying to somehow

please you, or put on a good face. There’s some pride issues; they’re not sure if they can

trust you. But I think over time, if you’ve got a good doctor-patient relationship, wow,

you may know things that only their priest or religious advisor knows. I mean, it can get

to that level” (P08, Internal Medicine). She cites a patient case in which the patient-

doctor relationship facilitated sensitive psychosocial information; “There are people who,

for reasons of pride, will really hide [financial barriers] … 20% of the time they’ll hide

it…. it may come out down the road. But only if I’ve established that relationship” (P08).

Quality of the Relationship Grants Access The quality of the relationship gives the physician access to sensitive, pertinent

psychosocial information. P08 also describes a case when the quality of the relationship

granted access to pertinent psychosocial information; “since I had a relationship [with

the patient] she knew that she could say that [she will not change certain dietary habits]

to me.” Another internal medicine physician (P09) describes how the quality of the

relationship can grant him access to the patient’s situation, important in understanding

level of social support; “… from [the] relationship comes all sorts of wonderful things,

including access to their social world…. Because they’ll start to be real with you after a

while. They may not tell you everything, but … I don’t tell my wife everything, either. You

know what I’m saying?” A family medicine physician (P16) with almost twenty years of

experience describes her mindset during her consultation with a new patient, and her

awareness that patients may not initially reveal pertinent, and very sensitive, psychosocial

information:

108

[Seeing a] patient for the first time, you have no idea. And, sometimes patients lie and say, “Everything’s fine,” even when they’re getting the crap beat out of them every night.… Or they’re doing a lot of drugs or they’re homeless…. They’ll lie because they’re embarrassed. But, usually, after a couple of visits, you can build some relationship and trust and you’ll start to get more information. So, there’s some continuity of care that leads to more disclosure. Continuity Physicians express the view that seeing the same patient enables them to “get to

know” them, and to some degree enables patients to “get to know” them. This continuity

is important to accessing psychosocial information; “over time, you get to know the

person. And it’s a chronic disease, so yeah … sometimes when you do see them every

month, there’s more things you find out and you can take it to account” (P05,

Endocrinology). Seeing the same patient over time helps facilitate disclosure which

occurs as the relationship develops, and trust is established:

I think trust plays into things a lot…. If we have an established relationship, so they also know me and they know that I’m not gonna force anything on them … that I care about them …. It’s gonna be a shared decision making versus someone who doesn’t know me at all … if I do know them, [if] I do know their family, and we kinda have gone through ups and downs together. (P15, Family Medicine). Trusted advisor Seeing patients over a period of time enables the physician get to know the patient

in depth. This helps the physician recognize if psychosocial factors may facilitate, or

present barriers to, following the recommended care regimen; “part of it is the

longitudinal rapport that I build up with patients. I mean if they’ve been my patient for

years, hopefully I am that trusted advisor that is giving them some salient information”

(P01, Family Medicine).

109

5.4.1. Techniques to Build and Maintain the Patient Relationship Physicians describe various techniques they use to help build and nurture trusting

relationships with their patients; accordingly, this is included in the cognitive map of

psychosocial information use (see Figure 5.2). This is an important part of how they

provide patient care. Building trusting relationships is at the core of how they practice.

“How” to Talk to Patients Techniques to build trust include how physicians talk with patients. They ask

general and specific questions that elicit the patient’s input on barriers to care that they

may be experiencing. They may direct the conversation toward specific barriers (i.e.,

financial strain) or guide the consultation toward how their patients’ living situations

might impact their self-care behavior, and subsequently their HbA1c; “I am … able to

bring the attention and the focus … on [what is impacting them]. Some of the times …

they’ll volunteer, ‘Listen, my control’s horrible because right now I’m feeling horrible

because of my grandson.’… Hopefully I’ll lead them in that direction” (P01, Family

Medicine).

Demonstrate Caring Successfully demonstrating caring for their patients is central to how physicians

deliver care. An internal medicine physician (P09) describes how caring is really the

essence of what he does with his patients, “I really don’t ‘cure’ very many people. I just

kind of care about them. And that caring translates to a relationship.” Demonstrating

that they care for their patients helps them forge connections, which facilitates sharing of

pertinent, often sensitive, psychosocial information; “[they disclose because of the]

110

doctor-patient relationship, the ability to make a connection … The fact that I care about

them and that I demonstrate that in my behaviors” (P08, Internal Medicine).

Safe Environment Physicians express that establishing and maintaining a safe environment – as

defined by freedom from judgment, affirming language, and protection of privacy — is

important. A safe environment is imperative to establish from the very first encounter

with the patient; “I think the key is, if you’re judgmental or blaming in the initial visits,

and you say, ‘What? You aren’t taking care of yourself.’ Or, ‘I can’t help you if you don’t

take care of yourself,’ then, you get nothing. But, if you’re supportive and say, ‘Wow, it

sounds like you’re working really hard to manage this,’ or, ‘You’ve done a good job of

dealing with this problem,’ then you get more information” (P16, Family Medicine).

Furthermore, maintaining this protected space facilitates access to information; “I’ll ask

one question and then this whole other floodgate opens … if you allow that possibility

and allow people to feel safe, which I think is really what it is” (P15, Family Medicine).

However, physicians must maintain appropriate boundaries if they have close

relationships with patients and their patient’s family members, whom they may also be

treating. This is a consideration in maintaining, and communicating, a safe environment,

free from disclosure of personal information. Once patients realize that their

confidentiality is maintained, they may share additional information:

There are two issues. If … the doctor knows your family … and your friends … very well… sometimes that kind of becomes like, “Oh, hopefully that doesn’t get disclosed.” Usually it doesn’t, but that confidence I should have in you that, “Okay, my information is going to remain personal.” [the patient may] start disclosing … As a patient, if I’m improving in my health, and I’m seeing that my doctor is really doing the best to work on my health, and I haven’t seen anything getting disclosed anywhere, I’ll start disclosing accordingly. (P11, Internal Medicine)

111

Empowering Patients Physicians also attempt to develop and nurture the patient relationship by

acknowledging the autonomy and dominion that patients have over virtually every self-

care decision, even the decision to select their doctor. This means that some physicians

refuse to be directive with their patients. As this physician says, “I will not necessarily

tell a person what to do. Some people want that, then I’m not a good doctor for them. I’ll

just tell them straight to their face, ‘I’m just not gonna be a good doctor for you. There’s

lots of other doctors’” (P14, Family Medicine).

Empowering interactions are also characterized by decisions that are shared. An

endocrinologist (P05) describes how she empowers patients by allowing them to voice

opinions about potential treatment options, “[I give them] some time to talk … [I] ask

them, ‘What do you think?’ or ‘Can you do this?’ Instead of saying, ‘Do this. Do this. Do

this.’ I think that’s the main thing.” Similarly, this family medicine physician (P15)

empowers patients through shared decision making, “they know that I’m not gonna force

anything on them… I’m gonna make recommendations, but it’s gonna be shared decision

making.”

5.5. Triggers to Gathering and Using Psychosocial Information In this section, I answer my fourth research question. I describe the situations that

trigger consideration of psychosocial factors.

RQ4: In which situations are psychosocial factors considered? Although practitioners consistently indicate that psychosocial factors are

generally important influencers of diabetes clinical care decisions, there are specific

circumstances when psychosocial factors are particularly relevant, and situations when

112

they are less so. In this section, I outline the situations in which physicians may not

consider psychosocial factors, followed by the situations when they do.

Do not Consider Psychosocial Factors if Patient is Responding to Treatment Physicians may not consider psychosocial factors if the patient is doing well

clinically; “If a patient is already well-controlled, typically whenever we see him and his

A1c is good every time. We really don’t think much beyond that” (P04, Internal

Medicine). Another internal medicine physician (P13) states, “I think that for better or

for worse, I often don’t consider … [psychosocial factors] … as much in my treatment

decisions … if everything is going smoothly.”

If the patient appears to be following the care regimen — and their clinical

numbers bear this out — the physician may conclude that, if indeed the patient is

experiencing barriers to care, they must be addressing them sufficiently; “You don’t

actually have to … [think about psychosocial factors] in the healthy guys that have it

together and have financial resources. They’ve figured out a way early on, when they

first got diagnosed, to fit diabetes management into their daily lives. They’re eating

healthy. They have a routine. It’s not a big deal for them” (P16, Family Medicine).

Another family medicine physician (P06) reiterates that he may not investigate potential

psychosocial barriers for the relatively few patients who are at goal, because the patient

must be addressing them effectively enough; “If they were always at 7 [HbA1c], then I

probably won’t go into a lot of depth about psychosocial barriers … there’s not all that

much we actually have to talk about; they just need refills, and that takes 2 minutes …

because … whatever barriers they may have … they’ve overcome [them]…. But [that is

not the case] for most patients.”

113

Situational Factors Next, I describe the situations that do trigger the consideration of psychosocial

factors, using data from the physician interviews (see Table 5.5), followed by the results

from the online survey (see Table 5.5.4). I segment these situational factors described in

the physician interviews into three areas: 1) chronic circumstances, based on ongoing

situations, 2) new circumstances, based on emerging situations, and 3) a change in

circumstance, based on a variation in situation.

Table 5.5 – Triggers Gathering and Using Psychosocial Information

Chronic Circumstance New Circumstance Change in Circumstance

At-Risk Patients – patients from groups the physician considers at higher risk to experience barriers to self-care due to psychosocial factors

New Patient – physician is seeing a patient for the first time

Spike in A1c – patient experiences a sudden elevation in HbA1c for reasons that are initially obscure

Patient not reaching goals on an ongoing basis – the patient is not reaching clinical goals (i.e., their numbers are “not what they should be”)

Initial T2DM Diagnosis – a patient that is diagnosed initially with diabetes

Sudden unhealthy self-care behavior – patient is abruptly not following the care recommendations, discovered through patient consultation, caregivers, or health record (medication refills, appointment no shows)

Chronic Circumstance I define chronic circumstance situational factors as those affecting patients with

persistent conditions over time. Physicians consider psychosocial factors when treating

patients who fit their definition of “at-risk patients” and those not reaching goals on an

ongoing basis. As shown in Table 5.5.1, physicians gather information such as financial

114

strain, access to healthy foods, and areas to exercise, to classify adherence barriers and

make referral decisions to help patients address them.

Table 5.5.1 – Chronic Circumstance

Psychosocial Trigger Information Gathered Use

• Financial strain • Mental health status • Payor status • Immigrant status

• Frailty • Psychosocial barriers

to performing activities of daily living

• Indicators of barriers to “taking care of themselves”

• Classification o At-risk o Complex

• Probe for self-care barriers

• Prompts further examination

• Analysis of the causes of patient behavior

Classifying At-Risk Patients – Complex Patients Physicians classify some patients as “complex”, based on their at-risk status

which can be due to their socioeconomic status, immigrant status, or mental health status.

Classifying patients as “at-risk” prompts physicians to routinely consider psychosocial

factors and potential barriers to self-care. An internal medicine physician (P08) explains

her thinking on how at-risk patients prompt her consideration of potential psychosocial

factors:

Psychosocial information is always going to be important, probably more important … in people who are frail, underserved, somehow at-risk, whether it’s … people with substance use … [people with] difficulty in performing independent activities of daily living … someone who is … very malnourished, recurrent falls, or other evidence that there’s some impairment in the ability to take care of themselves. Physicians consider impoverished patients, and those who experience financial

barriers, to be at-risk, and consider psychosocial factors categorically; “you always

consider them because if they tell you that there’s no way they can afford insulin or if

115

they tell you that it doesn’t fit in their lifestyle, then you know” (P03, Internal Medicine).

Poverty generally presents considerable barriers to care and has a cascading influence on

the demanding diabetes self-care regimen. One family medicine physician (P16) outlines

the interrelationships which categorically trigger her consideration of psychosocial

factors; “Extreme poverty, recent loss of a job, recently lost housing, house repossessed

or kicked out of a rental unit, serious mental illness, or physical illness that’s resulted in

significant disability and job loss.”

Physicians may determine that patients who directly communicate that they are

generally struggling with taking care of themselves may be at-risk, triggering

consideration of psychosocial factors. Physicians explore self-care barriers when they

classify patients as at-risk. Patients with diminished capacity, which can be caused by

various psychosocial factors, prompt further exploration of potential self-care barriers.

An internal medicine physician (P03) describes the characteristics that he considers

which put patients at-risk, and influences his priority of focus areas; “if you’re not

understanding, you have issues and that makes you a complex patient. You might be

depressed … wanting to die quickly. So that’s [what] we need to address.” A family

medicine physician (P12) reflected on his thirty years of experience with at-risk patients,

and their difficulty with navigating the diabetes care regimen:

I think the less sophisticated the person, the less success I have with complex [diabetes] regiments; number of pills, testing, multiple injections of insulin. It takes a very motivated, relatively sophisticated person to manage a complex health problem … I’ll try the best I can, but you have to individualize it. My experience is that [at-risk patients] have relatively little sophistication with regard to health concerns and health issues.

116

Immigrant status can serve as a trigger for assumptions, which can be imprecise.

An internal medicine physician (P09) discusses his experience with immigrant

populations and his indefinite approach to determining risk based on a patient’s

immigrant status. He reflected upon how his classifications may be imprecise, and reflect

his own bias:

I think poverty … In my community, I’ve cared for a lot of immigrant populations most of my career in the last 15 years or so. Not that all immigrants are impoverished, but immigrant status … the ability to navigate, the evidence that they’re navigating the culture that they’ve moved to, that they’re not socially isolated. We used to talk back in the day about linking social capital … So, you kind of do that … Filter through that grid of how nuclear are they, how tightly bonded they are to their immigrant group … Let’s say you came from a small town in Mexico, but you’re clearly navigating … Your temperament is one; that you are not shy. You’re really aggressive, and you’re not intimidated by the American healthcare culture or the American culture in general … I will have a less worry about … That really probably is a subconscious bias.

Patient not reaching goals, numbers “aren’t what they should be” Physicians consider psychosocial factors when a patient is not reaching clinical

goals on an ongoing basis, (e.g., when their HbA1c remains higher than goal). In these

circumstances, physicians consider psychosocial factors may be the cause, and probe

accordingly. An internal medicine physician (P13) describes the circumstance for his

patients; “It’s often when somebody’s HbA1C is 10 and they’re not taking their

medications … Then, we might start to ask some of those questions about … psychosocial

barriers … [things] we need to be aware of and … to address.” Physicians consider

psychosocial factors if a patient is not able to follow the treatment regimen, over a period

of time. An internal medicine physician (P03) describes when this could occur: “one

circumstance would be recurrent, non-compliance. Like, you’re not listening to what I

am telling you […] Then there might be a compounding factor.” This prompts physicians

117

to investigate potential barriers to care, which could be due to psychosocial factors.

Physicians attempt to determine if, and to what degree, psychosocial barriers may cause

persistent unhealthy self-care behavior. One internal medicine physician (P13) explains:

If their HbA1c is elevated, if their blood pressure is elevated beyond … the target.… If their cholesterol is elevated beyond what it should be and they are supposed to be taking a cholesterol-lowering medication that should, if taken correctly, at a proper dose, should be getting them to the target goal … Or they’re missing appointments. I think that those are the things that would prompt me then to say, sort of be a red flag … I should dig into some of those things a little bit deeper. Rather than an absolute number, an internal medicine physician (P07) speaks to

the situation in which the HbA1c and other clinical variables are not improving; “Their

blood pressure’s not under control. They’re on 4 medications … and they’re blood

pressure’s still not under control. Their A1c is not improving … they’re not responding.”

Another internal medicine physician (P09) expresses the circumstance more generally;

“When things are just whacked, you know, really out of control clinically. So that’s really

the driver to the social-behavioral, psychosocial factor consideration.”

In addition to the absolute numbers, knowing the trend of the clinical data is also

a trigger for consideration of psychosocial factors, resulting in conclusions about

psychosocial barriers to care. Following a patient’s progress over a period of time can

help physicians understand the efficacy of the regimen and discern if psychosocial factors

may be presenting barriers to care which result in an inability to reach clinical goals. A

family medicine physician (P01) states how he considers psychosocial information to

determine barriers if a patient is not reaching clinical goals; “I would say that those

patients that are not able to get to goal are ones that are going to get some increased

attention … there are lots of my patients that get that same exact intervention and still

118

have an A1c of 12.0. So, what’s different? That’s why I say …’Yeah, sometimes they

[psychosocial factors] matter a whole lot’… because it’s nothing to do with their

metabolism or anything like that. It’s what’s important to them and what other problems

are going on at home to cause them to not be able to follow the treatment plan.”

New Circumstance I define “new” circumstance as those in which there is a change for the patient or

for the physician. As shown in Table 5.5.2, physicians consider fairly distinct

psychosocial factors when treating patients new to the physician, and those patients with

an initial diabetes diagnosis. For new patients, physicians investigate level of social

support and self-care practices to determine what information exchange is necessary in

the initial consultation. For newly diagnosed patients, physicians gather information on

other chronic conditions and intention to make required lifestyle changes.

Table 5.5.2 –New Circumstance

Psychosocial Trigger Information Gathered Use

• New patient

• Social support • Background / intake

information concerning health behavior, self-care practices

• Potential barriers to self-care

• Assess causes of patient behavior (barriers / facilitators)

• Assess patient capabilities • Assess patient

understanding

• New diabetes diagnosis • Other chronic conditions

• Intention to make required lifestyle changes

• Assess causes of patient behavior (barriers / facilitators)

• Assess patient capabilities • Assess patient

understanding

119

New Patient Physicians consider psychosocial factors when seeing a new patient. The first visit

can include gathering psychosocial information about various specific psychosocial

factors, including level of support, physical activity and general health habits. An internal

medicine physician (P07) explains: “… in the first visit [I] get the information in terms of

what support they have and who’s at home.” Another internal medicine physician (P08)

explains how she asks about physical activity, and potential barriers; “[I ask new

patients] about their health habits and the things that affect those health habits … [I] ask

them about what they do for physical activity … [I] ask them about what are the things

that they’re doing to manage what makes it easier or harder [for them].” P07 also

explains her probes concerning mental health; “I will ask about mood as well … that

might give me a sense [of their psychosocial factors].”

Psychosocial information gathered in the initial visit is by no means

comprehensive, but it can be helpful for physicians to get an initial understanding of the

patient; “it’s not gonna be as deep a dive, but still getting at least some of those initial

factors” (P08, Internal Medicine). Physicians can also get background information on

new patients via intake questionnaires, which collect information on self-care habits, such

as dietary practices. As (P15, Family Medicine) notes, “[At the] first appointment … I

will have a lot of … information … [our] 25-paged intake packet … includes a diet diary

and [things] … that we think make a difference in people’s health … so we ask about it. I

think we have the SF-20 (20-Item Short Form Health Survey) … in there and … a lot of

different measures of just kind of health and well-being”

Physicians recognize that information that addresses the full range of

psychosocial barriers which could influence care decisions, cannot be gathered in the

120

initial visit. The volume of information concerning diabetes, and the practical constrains

on how much the patient is able to digest, dictate the extent to which physicians can

collect psychosocial information with a new patient. P10 (Family Medicine) explains:

“…it’s hard … I almost feel like we need 2 or 3 new patient visits because it ends up

being so much information to a patient … there’s only so much a patient’s gonna hear

from you, or really understand … sometimes [patients] come [directly] from the hospital

… they have a lot of new stuff going on that they didn’t even know they had until they

went to the hospital”

New Diabetes Diagnosis A new diabetes diagnosis may prompt physicians to consider psychosocial

factors. As P13 (Internal Medicine) said, “certainly, a new diagnosis of diabetes when

somebody is hearing this for the first time [prompts considering psychosocial

information]” This is particularly important if a patient has preexisting chronic

conditions; “they’ve had heart failure or a heart attack and diabetes and cholesterol and

it’s all new to them” (P10, Family Medicine). Physicians recognize the scope of the

effect on the newly diagnosed patient; “a new diagnosis of diabetes … you think about

how that information’s gonna interact with somebody’s views of themselves, with their

environment … in some cases, [they’ll be getting] much more healthcare and be followed

much more closely than they had been used to before … especially if we’re having to start

them on insulin [right away]” (P13, Internal Medicine).

Change in Circumstance I define change in circumstance as when there is an obvious change in clinical

numbers, or sudden unhealthy self-care behavior that is inconsistent with the

recommended diabetes care regimen. As shown in Table 5.5.3, physicians consider

121

psychosocial factors in the following circumstances: a “spike” in HbA1c and sudden,

unhealthy self-care behavior. Physicians gather information on potential barriers to care,

such as financial strain and missed appointments.

Table 5.5.3 – Change in Circumstance

Psychosocial Trigger Information Gathered Use

• Spike in HbA1c • Sudden unhealthy self-

care behavior

• Sudden psychosocial barriers o Financial strain o Missed

appointments

• Analysis of the causes of patient behavior

• Assess needs arising from barriers to self-care

• Referral decisions

Spike in A1c Psychosocial factors are considered when the physician observes an abrupt

increase in HbA1c. Physicians specifically mention that a sudden increase in HbA1c (i.e.,

a “spike”) triggers consideration of psychosocial information; “so if you’ve seen the

patient, if it’s not the first visit, and you see their A1c spiked … it could be that trigger to

consider psychosocial factors” (P06, Family Medicine). Another family medicine

physician (P15) explains how she might probe for barriers to self-care when she observes

a sudden change in HbA1c; “I think that … either like a clinical change … a spike in A1c

… I’m probably gonna probe more into why I think we may be having difficulty.”

Another family physician (P01) explained how in these circumstances he may consider a

decision to get support for access to medications:

For those patients who either historically have had great successes with their sugar control, and then all of a sudden they’re having a quarter where their A1c jumps up … from a 7 to an 11 … in 1 quarter … almost to a person that’s either someone who has had a total lack of financial resources, so they were well-controlled on medication X and now they’re no longer able to get that medication.

122

Sudden unhealthy self-care behavior Psychosocial factors are explored if the patient is not able to achieve various

elements of self-care. Physicians consider psychosocial factors because they believe a

sudden change could stem from a patient not following the treatment regimen due to

various self-care barriers that have recently emerged due to shifting psychosocial factors.

A family medicine physician (P12) explains his rationale for considering psychosocial

factors; “I would say, if there’s a change, either a disease that was under a certain level

of control is now out of control.” Another family medicine physician (P12) describes

how a patient missing appointments prompts investigation regarding potential

psychosocial factors. A decision to enlist other members of the care team to help locate

these patients is based on what information he is able to gather. He shared his experiences

with diabetes patients who may experience sudden, unhealthy self-care behavior. This

type of sudden decline in healthy self-care behavior is usually a signal that the patient is

experiencing barriers to care based on psychosocial factors, resulting recommending care

managers try to locate these patients, “If they [diabetic patients] were really adherent,

and, all of a sudden, they disappeared off the face of the earth…. [some] would be using

… drugs … or were … in jail … but we did the jail clinic so that was pretty easy to figure

out. We would know that.... The next time I went to the jail, they would … be on my

patient list … so generally, if they disappear, we would have our case managers give

them a call or track them down.”

Survey Findings I now detail the survey findings concerning the triggers raised in gathering and

using psychosocial information. By a large margin, survey respondents most frequently

indicate that psychosocial factors are important to consider in all circumstances. As

123

shown in Table 5.7.4, a wide majority of responses across respondent roles indicate that

they consider them in all circumstances (as noted in the table, survey respondents could

indicate more than one source). In this way, survey results differed from the interview

results. Nonetheless, survey results demonstrate that for some practitioners, each of the

categories of circumstances (chronic, new, changed) serve as specific triggers to the

consideration of psychosocial factors.

Table 5.5.4 – Circumstances when Psychosocial Factors are Important to Consider

Total (n = 204)

Primary Care Physicians

(n = 41)

Nurse Practitioners & Diabetes Educators (n = 163)

In all circumstances 182 (44.2%) 35 (29.7%) 147 (50%) Chronic circumstances

Patient with multiple chronic conditions 34 (8.3%) 11 (9.3%) 23 (7.8%)

Patient with persistent, low treatment adherence 33 (8.0%) 11 (9.3%) 22 (7.5%)

Patient from low-resourced areas 31 (7.5%) 12 (10.2%) 19 (6.5%)

Patient with diagnosed mental health condition 29 (7.0%) 10 (8.5%) 19 (6.5%)

Patient with undiagnosed mental health issues 22 (5.3%) 9 (7.6%) 13 (4.4%)

New circumstances Seeing a new patient 33 (8.0%) 12 (10.2%) 21 (7.1%) Seeing a work-in patient 14 (3.4%) 6 (5.1%) 8 (2.7%)

Change in circumstances Change in health status (e.g. spike in HbA1c, additional diagnosis)

34 (8.3%) 12 (10.2%) 22 (7.5%)

TOTAL 412 118 294 Note. Respondents could indicate more than one circumstance.

124

5.6. Patient Assessment In using psychosocial information, practitioners seek to make several judgments

that then inform their clinical decisions. I now outline how psychosocial information is

used by physicians in their general assessment of patients. As shown in Table 5.6, these

judgments include: 1) the causes of patient behavior, 2) patients’ capabilities, in light of

psychosocial barriers and facilitators that they may be experiencing; and 3) patients’

understanding of their diabetes care regimen.

Table 5.6 – Psychosocial Information Use for Patient Assessment

Practitioner Assessment Type

Practitioner Judgments Relevant psychosocial factors considered

• Causes of patient behavior

• Barriers and facilitators experienced

• Life stressors • Social support • Health beliefs and

preferences

• Patient capabilities • “Can they do it?” • Mental health • Health literacy • Payor status • Financial strain • Life stressors • Social support

• Patient understanding • “Can they understand it?” • Health literacy

The causes of patient behavior Psychosocial information is crucial for the physician to understand patient

adherence practices, in particular, why the patient may not be following the diabetes care

recommendations. A full understanding of self-care behavior is reliant upon

understanding if the patient may be experiencing psychosocial barriers to care, such as

financial barriers which may intersect with other psychosocial factors. Practitioners

described using information regarding life stressors, mental health, and social support in

125

such causal analyses. For example, an internal medicine physician (P13) describes how

becoming aware of a change in a patient’s life stress and a loss of social support was

responsible for changes in his previous self-care behavior: “a gentleman … in his mid-

50s, poorly controlled … for a while … we were trying to escalate his medications and

trying to get him to take his insulin routinely … We didn’t realize that he had changed his

job recently to a 3rd shift. He’s had all kinds of social challenges. He previously was a

cocaine abuser. He stopped that. He had a girlfriend who was a healthcare provider …

he started [to] ... clean his act up ... but then they broke up … talking about the

psychosocial stuff … [the girlfriend] was a supporter in his care, but when they broke up

that … provided him, I think, with enough confidence … to do some of the stuff on his

own.” A family medicine physician (P15) also shares a patient case that helps illustrate

his use of psychosocial information to identify life stressors and poor mental health as

causes of worsening patient self-care behavior:

His diabetes was poorly controlled.… I knew he was very motivated, and I couldn’t figure out what was going on … he was … very highly functioning, worked really hard on his diet, almost obsessively … was exercising regularly … then his A1c … jumped to … almost 10 … I was trying to figure out what had happened … as we were talking, there was this whole complex story that unrolled … his wife had gotten diagnosed with cancer, so that was one layer … the next layer was… they have been struggling in their marriage … he had actually been planning on asking for a separation. And then … she got this cancer diagnosis. So then he felt like he couldn’t leave her.… Even though he came in to talk about his diabetes, we ended up talking about his depression because he was now … in a marriage that he didn’t want to be in, and caring for this woman who had this very aggressive cancer diagnosis…. He was like, “I can’t leave her now.”… he both has the caregiver stress and the stress of [being in] a relationship that [he didn’t] want to be in anymore, and [he’s] really depressed.… That definitely changed the way we were talking about his diabetes … really shifted into a conversation about self-care, and how [can] we help [him care for himself] … then how does the diabetes care follow from that?... He said, “Well, I haven’t really been … eating as well … and I haven’t really been exercising.”

126

Patient capabilities A key assessment decision physicians make is to what degree patients are able to

manage themselves, their environment, and their health conditions. The ability to

“manage their lives” (P01) is an important assessment. One family medicine physician

(P16) explains that the assessment is based on how the patient responds to her questions,

“… [answers] give you an understanding of how much control they have over their lives

and how routine their lives are, and [any] stressors there are” (P16, Family Medicine).

Patient understanding A patient’s ability to understand their illness and its treatment is also an important

assessment that physicians make using psychosocial information. For example, a family

medicine physician (P16) with almost twenty years of experience in various care settings

speaks to the impact of how extreme cognitive impairment can influence self-care in

general; “It’s complicated to manage diabetes, but if you’re cognitively intact and have

some resources, it’s doable, even with other illnesses … [If the patient is] either very sick

or very cognitively impaired … you can’t really do it.” Physicians consider mental health

issues in order to gauge a patient’s ability to follow particular aspects of the care regimen,

as not following recommendations could put them at considerable clinical risk:

It might be dangerous for the patient if we’re doing something like administering insulin or having them take an oral medication that will drop their blood sugar. [If] they are going through a manic episode or they … have active psychosis. At best, we’re not gonna be doing anything with their treatment. At worst, we’re hurting the patient with that treatment. So, in that situation, there are some things that take a higher priority and that very clearly affects that decision making. (P07, Internal Medicine) A patient’s health beliefs and perceptions about treatment options, especially how

insulin and using needles impact their preferences, is also a factor relevant to patient

understanding. Some patients believe that administering insulin is uncomfortable, or

127

painful. One physician explains, “injecting could be up to 4 times a day … it’s a fuss,

they don’t want to be in pain” (P04, Internal Medicine). A family medicine physician

adds, “a lot of them really are afraid of the needle” (P12, Family Medicine). Some

patients are reluctant to start insulin because they believe that insulin may make them

sick or even kill them. Another family medicine physician (P06) shares his patient

experience, “for some patients, there was the fear of the association of death with

insulin.”

These beliefs primarily emanate from the patient’s interactions with and

observations of the experiences of those close to them—friends or family members with

diabetes who may have had to take insulin. Physicians from various care settings

discussed their experiences with patients with these beliefs about insulin and the use of

needles. As one family medicine physician (P12) shared, “a lot of them have family

members or friends who are on insulin. And more often than not, they’re varied

experiences, but a lot of them had negative feedback from other people … like they don’t

like injecting themselves….” As P10 asserts, “There’s lots of misconceptions about it….

A lot of patients would always tell me, ‘When you get insulin you get really sick.’

Physicians noted that these pre-existing, negative beliefs and associated

misunderstandings are important to identify since they could stand in the way of optimal

treatment for a given patient. As such, practitioners consider health beliefs in the context

of their overall assessment of patient understanding.

Psychosocial information helps to inform physicians’ assessments of patient

understanding of their diabetes care regimen, which helps them forecast the self-care

barriers they may confront.

128

In sum, psychosocial factors play a key role in informing the physician’s overall

assessment of a patient, as they attempt to recommend a treatment regimen appropriate

for the patient. It is to this issue that I now turn.

5.7. Making the Clinical Decision As shown on the cognitive map (see Figure 5.2), psychosocial information helps

physicians determine what treatment is most appropriate for the specific patient, at that

particular time. For example, a patient’s social circumstances can change quite

frequently, and assessing their overall situation is done, at some level, during each visit.

Psychosocial information, gleaned primarily from the patient, influences the physician’s

clinical judgment concerning what regimen is best, informed by their assessment of what

may be currently effective. I now detail how psychosocial information is used for four

types of diabetes care clinical decisions: 1) medication management, 2)

recommendations, 3) determining target level of control, and 4) referrals.

5.7.1. Medication Management Psychosocial information helps to inform medication management decisions. As

shown in Table 5.7.1, physicians use information such as payor status, financial strain,

social support, health beliefs, and occupational demands to inform judgment concerning

how well a patient is “managing their life” which helps determine the medication

regimen appropriate for the patient’s circumstances. Medication management decisions

include starting a patient on insulin, and determining which type of insulin is appropriate.

129

Table 5.7.1 – Psychosocial Information Use for Medication Management

Practitioner Judgments How Used Medications Decisions

• Barriers and facilitators experienced

• Determine feasibility of medication options

• Determine needs emerging from barriers to medication adherence

• Select medication type • Start medication therapy

(i.e., pills, insulin)

• “Can they do it?” • Determine feasibility of medication options

• Assess clinical risk of medication options

• Calibrate medication regimen complexity and intensity

• “Can they understand it?”

• Determine feasibility of medication options

• Determine needs for patient education

• Educate patient

Select medication type Insight into the barriers experienced by patients, and their related capabilities,

plays a role in physicians’ judgments about the feasibility of different medication options

for their patients. In turn, this leads physicians to prescribe medications that were deemed

to be most feasible from the point of view of patient adherence. For example, health care

payor status (closely associated with financial strain) means that some physicians actively

consider medication cost when selecting medications for patients, thus making

medication decisions according to a patient’s ability to pay. One family medicine

physician (P10) who practices in an urban, community clinic explains:

…we’re very aware of how much medicines are. We either get medicines that are $4 at Wal-Mart or that are discounted … or we get it through an assistance

130

program. We’re not just gonna write you random things that we know you’re not gonna be able to purchase … Because then you’re not completing the treatment that we’re recommending…. We make sure that we only recommend medications that a patient’s gonna be able to obtain. We’re not just gonna write it [just to] write it. We want the patients to actually take the treatment. A major medication type decision affected by payor status as a barrier also

concerns the type of insulin prescribed: specifically, fast-acting or slow-acting insulin.

An internal medicine resident (P04) shares, “[I consider] what type of insurance would

need to cover it [medication]. Typically, there are two main regimens ... long and short

acting … the longest one is quite expensive … most insurance companies cover it by now,

but there are some instances where … it’s not covered.… And of course all Medicaid

policies aren’t the same…. So that’s another thing that goes into our consideration.…

Cost would be a major … psychosocial factor … the ability to pay [for certain types of

insulin].”

Similarly, a family medicine physician (P10) practicing in a community clinic in

Texas shared how financial barriers due to lack of insurance coverage determine the

insulin type available to some of her patients, because of their immigration status and

lack of access to prescription assistance programs:

The Levemir and Lantus … the 24-hour insulins … are very expensive … the slow-acting … NPH [Neutral Protamine Hagedorn] 12-hour insulin … is cheap … a lot of our patients are on [Levemir and Lantus]. The way we get them is through assistance programs that the manufactures provide, but they are now requiring a social security number for these applications ... they used to accept a tax ID number [so if patients] had a tax ID, they could apply and get assistance. Well, they’re rejecting those now…. If they don’t have a social security [number], which a lot of our patients that are undocumented [don’t have, they] can’t qualify, can’t receive those, and so we’re stuck using older insulins, that are harder to manage, and harder to titrate into … If you give it too strong, it drops them too low. If you don’t give it strong enough, they don’t get low enough. And it’s a little bit more complicated … all based on the patient, being undocumented, not having a social security number, not having health insurance, and not being able to get their insulin.

131

The cost of fast-acting insulin also causes physicians to select the slower-acting

version of insulin in cases of general financial strain as opposed to just payor status. An

internal medicine physician (P03) explains how out-of-pocket medication costs factor

into his medication management decisions, which may be suboptimal clinically, because

of financial considerations:

Ideally you would prefer for them to be on a short acting and a long acting … Like TIDASE 3 times daily before meals and a long acting Basal. But then that’s … a lot more expensive than an intermediate acting insulin.… Many times they just come tell you, “I can’t afford that insulin.” [Because they] can’t afford it…. [They run] out of insulin… So, a lot of people have to be on 70-30 [intermediate acting insulin] … mainly because of cost reasons … which is not really ideal.

Start medication therapy Start Pills The physician’s assessment of a patient’s capabilities and barriers/facilitators is

considered in light of treatment options that typically occur before medication therapy:

lifestyle changes in relation to diet and exercise. Specifically, if a practitioner does not

think a patient can make the requisite changes, more aggressive treatment options may be

initiated sooner. For instance, an internal medicine physician (P13) describes a

representative patient case in which he makes a decision to start medication based on his

predictions regarding a patient’s future self-care behavior based on his/her financial

barriers:

so if I have a patient, for example, who is not as well-controlled as they could or should be with diet and exercise alone, and they have low socioeconomic status, live in a low income neighborhood, where it’s not safe for them to exercise outside, it’s difficult for them to have affordable access to healthy food, I might be less confident that they’re really gonna be able to get by on diet and exercise alone. In that case ... we may need to go straight to medication.

132

Start insulin The decision to start insulin is one of the key medication management decisions

that requires judgment informed by psychosocial factors. As such, I now describe the

psychosocial factors which specifically influence the decision to start insulin. Patient

understanding, as manifested through their health beliefs, may also influence the decision

to educate patients about adding insulin to their treatment. Physicians note that adding

insulin to the diabetes care regimen increases its complexity. An internal medicine

physician (P13) acknowledges what a big adjustment adding insulin to a medication

regimen means for some patients; “For some people, it means they’re gonna have to take

medication every day for pretty much the rest of their lives. When they’re not used to

doing that … [it] could be a big gear-shift. So I think about those things very much in

that situation [starting insulin].”

Health Beliefs and Preferences Physicians explain how strong health beliefs are important factors to understand

and consider as they deliberate over the decision to start insulin. They attempt to

understand these concerns and when they encounter patients with intense fear of insulin,

they determine the patient’s needs for patient education. Following this, they attempt to

educate the patient on what starting insulin might mean for them, while addressing their

misconceptions: “… I’m like, ‘No, it’s because you’re really sick with diabetes that you

need insulin. It’s the high sugar that’s causing this, not the insulin’” (P10, Family

Medicine). Physicians often have discussions with their patients concerning these health

beliefs and preferences over a period of time, as the physician educates the patient, and

they negotiate the significant decision to add insulin to their care regimen. Notably,

efforts to persuade are woven into efforts to educate. A family medicine physician (P10)

133

shares, “[what] happens … a lot is we have patients that are very resistant to insulin and

they would do much better if they took insulin early on … we have a hard time … we’ll

spend several months trying to convince someone that they’re gonna need insulin

because they’re just not getting anywhere on maximum doses of oral medications. And

then once they finally get on insulin, they’re able to see some improvement.” An internal

medicine physician (P12) explains that he discusses insulin with patients well before it is

clinically necessary to include it in their care regimen. His rationale is to quell the health

belief that diabetes is a simple disease to treat, one that may just require a pill. He makes

the decision to educate diabetes patients early on; “I throw the insulin concept out onto

the table very, very early on, so they kind of understand what the disease is. Because a lot

of people [believe], ‘All I have to do is take a pill.’ And then think they can get off the pill

… Like you take an antibiotic for strep throat, you take this for a period of time, then I

can come off it. Or that once their glyco is getting better, ‘now I can come off the

insulin.’… I’ve been through this path before, so I start the discussions earlier on.”

Despite their efforts to educate, physicians may not start patients on insulin until

long past the time when it could benefit them. This is because some patients do not

believe they are sick, or sick enough, to need insulin. They perceive the additional

responsibilities as considerable inconvenience to their care regimen; they surmise that

insulin is not worth the benefit. Patients may not “feel” sick, therefore they do not believe

or fully understand the implications of chronic high blood glucose. It can be difficult to

convince asymptomatic patients to take on the considerable life changes required from

adding insulin to the care regimen; “I had many more patients who were not at 7, or

probably not even at 8, and needed to start insulin, but they really didn’t want to. [It is]

134

hard to convince the patient to start insulin, to use a daily injection and check their

sugars, to poke their finger several times a day, as well as inject themselves with the

medication, because their blood sugar … they can’t see and they can’t feel … when their

blood is abnormal … [it] puts them at risk for a health event they’ve not yet experienced”

(P06, Family Medicine).

Social Support The vital decision to start insulin requires physicians to assess feasibility and

clinical risk in relation to the patient’s level of social support. They approach this

decision thoughtfully and seek a full understanding of the benefits and risks based upon

level of social support; “you really have to be careful … especially if they don’t have

assistance at home … some people will have poor vision because of glaucoma, [or]

cataracts. And they’re not reliable to give themselves insulin. They’re not reliable to

follow a glucometer … because they might mis-read how much they’re getting and

[because] they can’t see well.… I’ve had a couple of patients like that … I definitely have

not pushed insulin aggressively … for their own safety … because it’s just not safe” (P10,

Family Medicine).

Life stressors Employment responsibilities are a life stressor that physicians consider in

assessing the feasibility of starting insulin. A family medicine physician (P17) shares

how he considers employment in making this decision. Patients who work in certain

occupations may experience barriers to adhering to the medication regimen which

includes insulin injections; “if I start someone on a regular insulin, getting insulin shots

3 times a day and getting your Accu-Cheks (glucometer) 3 times a day … [it] is just

difficult for a working-class patient. Where are they gonna put their pen and needles? It’s

135

just impossible. I won’t say impossible. It’s difficult. Let me put it this way … you’ll see

less compliance on those patients or less results outcome, because they may miss the

afternoon dose.”

Calibrate medication regimen complexity and intensity The assessment of whether a patient “can do it” is important in relation to

decisions about the clinical risk associated with complex or intense medication regimens,

of which regimens including insulin may be an example. In particular, the barrier of a

lack of social support is evaluated in relation to the clinical risk of an intense and

complex treatment regimen; physicians consider the number and type of medications

prescribed to the patient in relation to support with medication behavior. An internal

medicine physician (P08) based in Michigan describes a patient case in which low social

support influenced medication selection —due to the risk of episodes of low blood sugar;

“there’s nobody to check on him, nothing. So, my goals for him were actually convincing

him that, we needed to back off on the medications that his prior physician had been

giving him.”

In some cases, treatment intensity may also be lowered based on concerns about

the harms associated with a high risk of non-adherence. Indeed, some practitioners even

speak about suspending treatment in situations of mental health challenges. As this

physician said, “[If the patient is] either very sick or very cognitively impaired … you

can’t really do it.” Additionally, P07 notes that patients not following recommendations

due to active psychosis could put them at considerable clinical risk:

It might be dangerous for the patient if we’re doing something like administering insulin or having them take an oral medication that will drop their blood sugar. [If] they are going through a manic episode or they … have active psychosis. At best, we’re not gonna be doing anything with their treatment. At worst, we’re

136

hurting the patient with that treatment. So, in that situation, there are some things that take a higher priority and that very clearly affects that decision making. (P07, Internal Medicine)

Health Literacy Health literacy is another factor considered in determining appropriate regimen

complexity. An internal medicine physician (P13) describes how he may establish a

higher HbA1c target based on a patient’s ability to comprehend required self-care

behavior required, “if I feel like the person does not have a sufficient level of

understanding of how to take the medication, why we're doing it, how to do it safely ...

then we may go with a less aggressive, safer route for them … the oral versus injectable

medication is one example of that sort of thing… That will still be better than no

treatment at all.” A family medicine physician (P10) also considers health literacy and

patient safety to decide on a higher HbA1c target, “some patients don't always know how

to even read their glucometer… you really gotta be careful on how aggressive you are

with insulin. If you risk them dropping their sugar too low, they don't even know how to

read or record the glucose … that could be really dangerous for them.”

5.7.2. Recommendations Psychosocial information helps to inform recommendation decisions because

physicians use this information to make judgments regarding patients’ behavior, their

capabilities, and the barriers/facilitators that they face. As shown in Table 5.7.2,

physicians use information such as financial strain and employment demands to perform

pattern matching described in chapter two (section 2.4). A family medicine physician

(P15) indicates, “If I know that somebody [is] from a lower socioeconomic status … that

… is gonna be much closer to the front of my mind and … the conversations that we’re

137

having … [and] the recommendations I’m making.” Further, they use psychosocial

information to advise, educate, counsel and determine the feasibility of the patient

following the care regimen. Physicians try to incorporate psychosocial factors into their

care recommendations so that their suggestions are practical (i.e., within the patient’s

reach of performing). In this section, I describe the how psychosocial information is used,

according to the following specific recommendations decisions: frequency of follow-up

visits, exercise, dietary and lifestyle recommendations, and supporting patients to

prioritize their own self-care.

Table 5.7.2 – Psychosocial Information Use for Recommendations

Practitioner Judgments How Used Recommendation Decisions

• Barriers and facilitators experienced

• Determine needs emerging from barriers

• Adjusting frequency and timing of follow-up appointments

• Adjusting exercise, dietary behavior recommendations

• “Can they do it?” • Determine feasibility of recommendations

• Acknowledging patient autonomy

• “Can they understand it?”

• Determine needs emerging from patient’s level of understanding

• Supporting patients to prioritize self-care (e.g., exercise)

Frequency and Timing of Follow-up Visits At a high level, diabetes care appointments can occur at fairly regular intervals,

however patient circumstances based on psychosocial factors can dictate how often

follow-up appointments are necessary, and the timing of those appointments. To

determine whether such adjustments are warranted, physicians attempt to determine a

patient’s needs as arising from the psychosocial barriers and facilitators that they

138

experience. For example, patients may use their previous experiences with patients in

similar circumstances to determine these needs. An internal medicine physician (P09)

shares how he matches the characteristics of current patients to patterns from previous

patients, which can inform the decision for frequency of follow-up visits:

I might ask a patient to come back and visit me more frequently than I would another patient. And that’s where this profiling issue really comes up. I look at you and my pattern says, ‘Oh, you’re one of those kind of folk. You’re low income. You’re from a racial minority group. You’re the matriarch of the family.’… Something triggers my thinking saying, ‘Oh, you’re gonna need a lot more support than if you were the male, diabetic, who goes to work, but doesn’t have to go shopping, doesn’t have to do the cooking, doesn’t have to do the cleaning.’ For example, when physicians perceive that patients may need support for

understanding care recommendations, they will attempt to schedule clinical visits when

family members can also attend. A family medicine physician (P01) who sees patients

from various parts of the world who have relocated to the Northeast part of the United

States, shares his approach to scheduling visits that may help patients who experience

transportation barriers to attending follow-up appointments: “…more than half of our

patients are non-English speaking. So, it’s a lot easier to book an interpreter to see me,

and to see the diabetic educator back-to-back, so we book … the van to come [and pick

them up]. The likelihood that a person’s gonna miss 1 visit versus 2 visits is less, so we

try and, we call them a tandem visit; just back-to-back.”

Adjusting Dietary Recommendations Nutritional guidance and dietary recommendations are influenced by psychosocial

factors, such as financial barriers, culture, employment and other lifestyle demands.

Physicians acknowledge patients’ autonomy concerning dietary choices, and attempt to

encourage them to make healthy dietary choices; “What the patient can decide is how

139

much … they are gonna eat and what are they gonna eat and when are they gonna eat …

I can’t really decide but I can advise, I can suggest” (P16, Family Medicine). An internal

medicine physician (P07) explains how recommendations are somewhat limited by the

extent to which she can control patient’s eating behavior, contrasting dietary

recommendations with medication prescriptions, “food ... It’s not like a medication. You

can’t really prescribe food, right, because everyone has their own preferences….

understanding their financial situation as it pertains to food helps.”

One way of tailoring dietary recommendations is to recommend specific foods,

advice that is adapted to psychosocial factors such as financial barriers. As P08 explains,

recommendations are made while: “…making sure of something he could eat …

something that he could afford, because he was on a fixed income … psychosocial

circumstances [have] a huge impact on how I worked with them for their diabetes” (P08,

Internal Medicine).

Specific dietary recommendations may also be made in relation to a patient’s

eating schedule. For example, these recommendations may be informed by a patient’s life

stressors: “…stress they have in their work…. People travel a lot … depending on …

their job duties. We probably make decisions in terms of what kind of meals they can

have” (P11, Internal Medicine). Furthermore, employment demands may guide

physicians toward recommendations aimed at educating and helping them address

barriers, in order to support them in their efforts to follow the diabetes dietary regimen:

Depending on … their job duties … we probably make decisions in terms of what kind of meals they can have, where they are, when they buy, when they cook … How many breaks they can take … we have to kind of educate them really well, “even though I understand that you are really busy … in spite of that, you need to take care … to maintain your disease … if you do these lifestyle modifications it helps maintain your blood sugar levels.” (P11, Internal Medicine)

140

Supporting Patients to Prioritize Self-Care In relation to recommendations, practitioners may also opt to support patients to

prioritize their own self-care in light of competing responsibilities, such as patient

caregiver responsibilities. Physicians attempt to encourage patients to not neglect

themselves when confronted with lifestyle demands; “[I tell patients with significant

caregiver responsibilities] ‘you need to take care of yourself. Is there a way that, on a

daily basis, you can still remember to do the things you need to?’ [I] remind them, ‘You

gotta take care of yourself in order to take care of other people’” (P05, Endocrinology).

An internal medicine physician (P08) shares how she attempted to support a patient with

significant caregiver responsibilities; “I wanted to work with her … first focusing on diet

and exercise, because I knew she was serving as a caregiver for her husband who was

blind, and was spending a lot of time with him. So, [I] talk[ed] with her about how to

carve out some time to step away from her caregiver duties so that she would be able to

… exercise and some of these other things.”

5.7.3. Determining Target Level of Control Psychosocial information helps to inform decisions about target levels of control.

These decisions can be made at each clinical visit based upon patient circumstances

driven by psychosocial factors. They can be revisited and adjusted, based on the extent to

which the patient has reached a prior HbA1c goal. Psychosocial factors influence the

setting of appropriate targets for control because the results of a risk assessment may

mean that the objective of treatment may be trying to reduce the risk of hypoglycemic

episodes, rather than attempt to create a regimen better suited for more ideal

circumstances. As shown in Table 5.7.3, physicians use their judgments about

barriers/facilitators, patient capabilities and understanding to help assess risk and

141

determine feasibility of “tight” control for HbA1c (i.e., HbA1c of 7). Patient preferences,

based on how the patient “feels” also inform appropriate target levels.

Table 5.7.3 – Psychosocial Information Use for Target Levels of Control

Practitioner Judgments How Used Target Decisions

• Barriers and facilitators • ”Can they do it?” • “Can they understand

it?”

• Assessing clinical risk

• Determining feasibility of different levels of control

• Choose HbA1c target • Acknowledging patient

decision-making autonomy

Choose HbA1c Target Establishing an appropriate target may also involve intersecting decisions,

determining risk and feasibility based on psychosocial information. For example, an

internal medicine physician (P09) states how assessment of both factors inform the

choice of HbA1c target: “…people who have erratic eating schedules, such as homeless

people, or people that have really, really strict incomes, and their social environment’s

really unstable … [I’ll have] them be less than 8 rather than less than 7 as a control

point on their A1c.” I now discuss issues of risk and feasibility in greater depth.

Assess risk Patient capabilities are assessed in relation to clinical risk. Risk of low blood

sugar is a vital concern for patients who lack social support, thus spending considerable

time alone. For example, a HbA1c target above seven may be appropriate, given their

circumstance. A family medicine physician (P14) discusses how demands, exacerbated

by low social support, drive her rationale for diverging from establishing strict control

goals; “somebody who had, maybe a lot going on in their life, and other issues…. I would

continue to work with them about their diabetes but, [I would not decide for] intense

control.” A family medicine physician (P16) describes how she determines appropriate

142

control targets based on financial strain; “If somebody doesn’t have a refrigerator or

regular place to live, electricity, regular meals … We would never tightly control this

guy. He would die…. I’d rather [him] have a HbA1c of 14 than try and manage insulin....

It’s not good, but the alternative is killing them. So, that’s worse, right?” An internal

medicine physician (P08) shares a case in which a patient’s low level of social support

had considerable influence on establishing an appropriate HbA1c goal:

He would go to Meijer’s [regional supermarket chain based in the Midwestern United States] everyday, because that was his major way of having contact with humans … So … I’m not gonna try to get him under a HbA1c of 7, because I think he’s at really high risk…. if he did get hypoglycemia at home, he would probably die, to be perfectly honest … because there’s nobody to check on him, nothing. Determine feasibility Physicians attempt to determine feasibility of a patient following a specific care

regimen required to reach a certain HbA1c target. Financial strain can have considerable

influence on establishing a target outside the guidelines. Physicians may explicitly divert

from known guidelines due to the patient’s financial situation: “…typically I would

consider them [financial barriers] all the time. So, when a patient presents I have to think

about the guidelines. Of course, there’s always sometimes where you need to break them

[financial barriers which dictate certain insulin which influences A1c target]” (P04,

Internal Medicine).

Patient preferences also influence feasibility of specific targets. Physicians gauge

the applicability of the guidelines, and may consider them less or intentionally disregard

them, based on patient preferences. An internal medicine physician (P03) expresses how

important understanding patient preferences are. He recognizes that ultimately, for

outpatient diabetes care, the patient retains final authority and autonomy over the vast

number of recommended self-care decisions. The patient makes decisions for dietary

143

practices, medication behavior, and attendance at clinical appointments — key

dimensions of the care regimen — regardless of physician’s objectives for them;

“patients can refuse treatment, right? They can take whatever [medications] they want.

You can’t really force it, right?” Patient preferences are a key factor in the decision to

establish an individualized HbA1c target. A family medicine physician (P01) emphasizes

how patient preference may supersede the guideline, or what he may believe is

appropriate for the patient:

[We must] make sure we’re assessing both what’s important to the patient as well as what we feel is important based upon evidence-based medicine. Because it’s one thing to come down from the mountain saying, “Well these numbers are important.” But again, if you can’t partner with them to say, why it’s important, or why it’s specifically important to them, I don’t think you’re gonna get a lot of traction. Acknowledging patient decision making autonomy An internal medicine physician (P08) described a patient case where a target was

established based on patient preference: “I don’t know if it’s a phobia, but he essentially

refuses to take medications. So he’s falling outside the [HbA1c] guideline.” Another

internal medicine physician (P03) describes how patient preferences, based on how they

feel, informs the decision to establish a specific, individualized target above the HbA1c

guideline; “say ... their HbA1c as 8, right? So, a person comes in saying, ‘Doc, this is

where I feel comfortable, I like my blood sugar at 140, I don’t like it when it’s 80, I just

don’t feel strong enough, right? I just don’t have the energy to do what I need to do with

life, I don’t care what the guidelines say.’ … then you just document it and let them be

happy because people know their body more than you know it, right?”

144

5.7.4. Referrals Physicians make referral decisions to help assess and get support for self-care

barriers that cannot be addressed during the clinical consultation. Referral decisions are

primarily driven by judgments regarding psychosocial factors that present barriers to self-

care. Referrals are made for counseling, behavioral health, and social work. Psychosocial

information helps to inform these referral decisions. As shown in Table 5.7.4, physicians

use psychosocial information to help determine barriers to self-care, and the degree to

which a patient understands the treatment regimen. Health literacy, financial strain, and

access to healthy foods and places to exercise help inform judgment on potential barriers

to care, and what steps to take to help address them. Practitioners make referrals to help

address barriers to comprehension of the treatment regimen, and support with dietary and

physical activity recommendations. A family medicine physician (P14) describes her

referral decisions quite simply, “I just listen to what they say …. then I … connect them

with [needed] resources.”

Table 5.7.4 – Psychosocial Information Use for Referral Decisions

Practitioner judgments How Used Referral Decisions

• Barriers and facilitators experienced

• “Can they understand it?”

• Determine needs emerging from barriers to self-care

• Determine needs emerging from barriers to understanding

• Referrals to assess barriers • Referrals to address

barriers

Referrals to assess barriers

Physicians make referrals based on assessments that are critical to supporting

patients in addressing barriers to self-care. One physician describes how she refers

145

patients to other members on her care delivery team to determine barriers to self-care,

“…I would refer them … either my nurse for case management or … to the pharmacist …

who really helps us a lot with our … patients … they’ll basically have a ½ hour phone

call with them every other week … [to] find out how they’re doing, what barriers they

have towards taking medications and things” (P07, Internal Medicine). An internal

medicine physician (P07) also describes how she refers patients to counseling staff to try

to identify, or confirm, barriers to care; “for me often times it’s trying to identity a

resource that I’ve got at the health center. Be it either a social worker, because the

patient’s having a problem with access to medications, or one of the behavior health staff

who can do some counseling, all the way up to getting them into see a mental health

professional like a psychiatrist. So, it’s trying to make a … fairly quick assessment about

what’s going on.”

Physicians also make referral decisions based on the patient’s level of health

literacy. These referral decisions are typically within the care team. They refer patients

with low health literacy to help assess their level of health literacy and to determine

whether a patients with low health literacy needs additional education.

Referrals to address barriers

Referrals outside of the care team One family medicine physician (P15) shares a specific patient case of a patient

who was experiencing mental health issues, “getting him into counseling, talking about

the depression, if we didn’t address this, his diabetes wasn’t gonna get any better.” One

family medicine physician (P17) states how he refers patients with mental health issues;

“if they are [depressed] … we look at the behavioral part and sometimes patients who

146

are having very high blood sugar or A1c [is] still not getting controlled, I refer them out

to BHC (Behavioral Health Consultants).”

Physicians may also refer patients to outside support services to help with

financial barriers. Social workers can help address financial strain that poses barriers to

care. An internal medicine physician (P07) discusses how she refers patients to social

work based on financial barriers; “I do refer patients for social work who tell me they’re

not having enough money … I, not infrequently, put a referral in for a social work to talk

to them about ways to help them … [find resources] in the community that might help

them out.”

Physicians may also refer patients outside the care team, directly to community-

based resources when patients experience barriers to access healthy foods or places to

exercise. These barriers are associated with a patient’s neighborhood and community

setting. They may live in areas that do not permit ready access to foods or places to

exercise, presenting barriers to self-care. A family medicine physician (P15) shares how

she attempts to determine if lifestyle demands or financial barriers are causing certain

dietary practices. If so, she will refer the patient to assistance:

I see people all the time, they’re eating off the dollar menu regularly, and so I sort of have a conversation about, what’s involved in that? Is it because they just really like fast food and they don’t really wanna change? Is it because it’s super convenient and they’re working 3 jobs …? Is it the cost issue?... I would address those very differently…. if it’s really a cost issue, then I’m probably going to be bringing in social work and making sure, finding out, do they have food assistance, [if] they eligible for food assistance. [ensure they] know what other resources are available … And I think it’s often multi-factorial, it’s not just one [psychosocial factor].

147

Referrals within the care team Physicians may also enlist the support of other members of the care team to

address low health literacy. Physicians may make global referrals to other care team

members to address health literacy issues. For example, this internal medicine physician

(P13) refers a patient to help them understand their treatment:

“[If a patient] has a low level of education, doesn't really seem to understand what's going on … in terms of understanding his medications. I ask him, ‘What other medical problems do you have?’ He has no idea and he's got all these things documented in the chart from the previous visit. So, I might say, ‘This person probably has low health literacy. This is somebody that I need to have meet with our social worker, have our nurse case manager reach out to.” Physicians may also make referrals to other care team members to address health

literacy surrounding more specific, rather than global, challenges. Self-monitoring and

medication management are two of the more significant problems addressed with specific

referrals. An internal medicine physician (P02) describes a representative patient case

when he scheduled a nurse visit to help a patient with finger sticks, “he [didn’t] know

how to … stick his fingers… I [brought] him back [for] a nurse visit to teach him that.”

Physicians may also refer patients to pharmacists to educate patients and provide

resources to help them with medication self-care. A family medicine physician (P12)

recounted how he referred patients on multiple medications to pharmacists to help

identify barriers to medication self-care due to inadequate understanding of what was

required, “if they're on multiple medications, and I have concerns about adherence …. we

would call our pharmacy … our pharmacy was great, we had some PharmD's there who

would do pill boxes for two to three weeks … [they] would do their pill boxes for them,

and then they would teach them.” Additionally, a family medicine physician (P12) refers

patients to other practitioners to provide patients with more detailed instructions on

insulin, “I [refer patients to] meet with the diabetic nurse … to have her just go through

148

some teaching … to let [the patient] see what these needles look like, and let [the patient]

practice with water. Just do it for a month. Not every day, just try a few [days], or have

them inject insulin in a very low dose.”

5.7.5. Survey Findings – Influence on Clinical Decisions I now detail the findings from the survey concerning the influence of psychosocial

factors on clinical decisions. The survey findings are consistent with the interview

findings concerning the influence of psychosocial factors on clinical decisions.

Target level of control is the decision most frequently influenced by psychosocial

factors, as shown in Table 5.7.5.1. Next in frequency are making recommendations,

followed by making referrals, and medications decisions. I isolate physician responses in

the table. The differences in responses between physicians and non-physicians for making

recommendations and other decisions are statistically significant, showing a difference

based on health care practitioner role.

Table 5.7.5.1 – Frequency with which Decisions are Influenced by Psychosocial Factors

Total (n = 157)

Primary Care Physicians

(n = 33)

Nurse Practitioners & Diabetes Educators (n = 124) p value

Target Level of Control 4.26 (.723)a 4.07 (781)b 4.31 (.703)c .113 Making Recommendations 4.18 (.777)d 3.83 (.879)e 4.26 (.729) .006

Making Referrals 4.13 (.830)i 3.93 (.712)j 4.17 (.808)k .219 Medications 4.09 (.722)l 3.98 (.649) 4.12 (.743)m .345

Note. Respondents were asked to indicate the frequency that the decisions they make, or have input into, are influenced by psychosocial factors. Responses were captured in a Likert scale: 5 – Always, 4 – Often, 3 – Sometimes, 2 – Rarely, 1 – Never. Standard deviations listed in parenthesis. a n=141. b n = 28. c n = 113. d n = 154. e n = 30. f n = 78. g n = 12. h n = 66. i n = 122. j n = 21. k n = 101. l n = 145. m n = 112.

149

Next, I highlight frequency with which psychosocial factors influence the specific

clinical decisions within each of the four groups.

Target Level of Control Decisions For target level of control decisions, psychosocial factors are most frequently

considered in incorporating input from the patient in setting the goal, as shown in Table

5.7.5.2. Next is establish target goal for blood glucose. I isolate physician responses in

the table. The differences in responses between physicians and non-physicians for

making incorporating input from the patient in setting the goal are statistically

significant.

Table 5.7.5.2 – Frequency with which Level of Control Decisions are Influenced by Psychosocial Factors

Total (n = 141)

Primary Care Physicians

(n = 28)

Nurse Practitioners & Diabetes Educators (n = 113) p value

Incorporate Input from Patient in Setting Goal 4.51 (.780) 4.15 (.970) 4.60 (.701) .024

Establish Target Goal for Blood Glucose 4.21 (.827)a 4.07 (.900) 4.24 (.808)b .333

Establish Target Goal for HbA1c 4.15 (.913)c 4.25 (.887) 4.13 (.921)b .519

Other Target Goal Decisions 4.15 (.950)d 3.82 (1.328)e 4.21 (.861)f .365

Note. Respondents were asked to indicate the frequency that the level of control decisions they make, or have input into, are influenced by psychosocial factors. Responses were captured in a Likert scale: 5 – Always, 4 – Often, 3 – Sometimes, 2 – Rarely, 1 – Never. Standard deviations listed in parenthesis. a n=140. b n = 112. c n = 140. d n = 68. e n = 11. f n = 57.

Making Recommendations Decisions For making recommendations decisions, psychosocial factors are most frequently

considered for dietary recommendations, as shown in Table 5.7.5.3. Making physical

acivitiy recommendations and recommending that the patient’s caregivers understand

what is required of the patient are next. Frequency of clinical visits and other

150

recommendations decisions are least frequently indicated as influencers. I isolate

physician responses in the table. I found statistically significant differences between

physicians and non-physicians in responses for making dietary recommendations, making

physical acivitiy recommendations, and other recommendations.

Table 5.7.5.3 – Frequency with which Recommendations Decisions are Influenced by Psychosocial Factors

Total (n = 154)

Primary Care Physicians

(n = 30)

Nurse Practitioners & Diabetes Educators (n = 124) p value

Make Dietary Recommendations 4.31 (.843) 3.90 (.995) 4.40 (.775) .003

Make Physical Activity Recommendations 4.18 (.961)a 3.66 (1.143)b 4.30 (.874) .007

Recommend Patient’s Caregivers Understand What Is Required of Patient

4.17 (.961)c 4.10 (.900)d 4.19 (.978)e .675

Frequency of Clinical Visits 4.12 (.912)f 3.87 (.937) 4.18 (.898)g .089

Other Recommendations Decisions 4.12 (.909)h 3.60 (.966)i 4.24 (.860)j .043

Note. Respondents were asked to indicate the frequency that the recommendations decisions they make, or have input into, are influenced by psychosocial factors. Responses were captured in a Likert scale: 5 – Always, 4 – Often, 3 – Sometimes, 2 – Rarely, 1 – Never. Standard deviations listed in parenthesis. a n=153. b n = 29. c n = 152. d n = 29. e n = 123. f n = 150. g n = 120. h n = 51. i n = 10. j n = 41.

Referrals Decisions For making referrals decisions, psychosocial factors are most frequently

considered in making referrals to a dietician and/or nutritional informaiton, as shown in

Table 5.7.5.4. I isolate physician responses in the table. None of the differences in

responses between physicians and non-physicians is statistically significant.

151

Table 5.7.5.4 – Frequency with which Referrals Decisions are Influenced by Psychosocial Factors

Total (n = 122)

Primary Care Physicians

(n = 21)

Nurse Practitioners & Diabetes Educators (n = 101) p value

Refer to Dietitian / Nutritional Information 4.26 (.980)a 4.00 (.775) 4.32 (1.016)b .122

Refer to Support Services Within the Organization

4.20 (.967)c 4.05 (.970)d 4.22 (.969)e .481

Refer to Diabetes Education 4.18 (1.024)f 3.90 (.889)g 4.24 (1.047)h .181

Refer to Support Services Outside the Organization

4.11 (1.027)i 3.95 (.805)j 4.14 (1.069)k .365

Refer to Specialty Care 4.07 (.873) 3.86 (.727) 4.11 (.898) .229 Other Referral Decision(s) 4.03 (1.025)l 3.43 (.976)m 4.15 (1.00)n .090

Note. Respondents were asked to indicate the frequency that the referral decisions they make, or have input into, are influenced by psychosocial factors. Responses were captured in a Likert scale: 5 – Always, 4 – Often, 3 – Sometimes, 2 – Rarely, 1 – Never. Standard deviations listed in parenthesis. a n=113. b n = 92. c n = 117. d n = 19. e n = 98. f n = 114. g n = 21. h n = 93. i n = 120. j n = 21. k n = 99. l n = 40. m n = 7 . n n = 33.

Medications Decisions For medications decisions, psychosocial factors are most frequently considered in

selecting a specific medication, as shown in Table 5.7.5.5. I isolate physician responses in

the table. None of the differences between physicians and non-physicians for the specific

medication decisions listed are statistically significant.

152

Table 5.7.5.5 – Frequency with which Medications Decisions are Influenced by Psychosocial Factors

Total (n = 145)

Primary Care Physicians

(n = 33)

Nurse Practitioners & Diabetes Educators (n = 112) p value

Select a Specific Medication 4.34 (.680) 4.21 (.600) 4.38 (.701) .173

Start a Patient on Non-Insulin Injectable Diabetes Medication

4.27 (.862)a 4.30 (.728) 4.26 (.908)b .798

Reduce Complexity of the Medication Regimen 4.26 (.822)c 4.09 (.843) 4.31 (.812)d .174

Select a brand, or a Generic, Medication 4.23 (.870)e 4.36 (.699) 4.18 (.919)f .300

Start a Patient on Injectable Insulin 4.21 (.862)g 3.97 (.695)h 4.28 (.897)i .075

Adjust Non-Insulin Injectable Diabetes Medication

4.03 (.987)j 3.85 (.972) 4.01 (.989)k .213

Add an Additional Oral Diabetes Medication 4.01 (.936)l 4.03 (.948)m 4.00 (.936)n .868

Adjust Insulin Injectable Diabetes Medication 3.96 (1.084)r 3.61 (1.298) 4.07 (.988)s .067

Start a Patient on 1st Oral Diabetes Medication 3.92 (1.079)t 3.81 (1.091)u 3.95 (1.078)v .536

Adjust Oral Diabetes Medication Dosage 3.87 (1.033)w 3.82 (1.074) 3.89 (1.024)x .722

Note. Respondents were asked to indicate the frequency that the medications decisions they make, or have input into, are influenced by psychosocial factors. Responses were captured in a Likert scale: 5 – Always, 4 – Often, 3 – Sometimes, 2 – Rarely, 1 – Never. Standard deviations listed in parenthesis. a n = 126. b n = 93. c n = 138. d n = 105. e n=132. f n = 99. g n = 136. h n = 32. i n = 104. j n = 124. k n = 91. l n = 130. m n = 31. n n = 99. o n = 53. p n = 10. q n = 104. r n = 137. s n = 104. t n = 130. u n = 32. v n = 98. w n = 135. x n = 102.

Practitioner Characteristics’ Association with Psychosocial Information Use Nurse practitioners, registered nurses, and diabetes educators indicate that each of

the four groups of clinical decisions are associated with psychosocial factors more

frequently than physicians indicate their influence. The difference is statistically

significant only for medications decisions. I found no statistical significance when

comparing experienced practitioners (>=10 years of experience) to less experienced

practitioners (<10 years of experience). Also, I found no statistically significant

153

difference in associations of psychosocial factors when comparing family medicine

physicians to internal medicine physicians.

Clinical Decisions Influenced by Psychosocial Factors and Years of Experience I also tested differences between frequency that psychosocial factors influenced

the four groups of clinical decisions, and years of experience. Years of experience is my

independent variable and each of the four groups of decisions is my dependent variable. I

measured the independent variable based on if the respondent had less than ten years of

experience, or greater than or equal to ten years of experience. I used the one-way

ANOVA to investigate difference. I found no statistically significant difference between

years of experience and importance of psychosocial factors for any of the four groups of

decisions (Recommendations: F(1,152) = 3.34, p > .05; Medications: F(1,143) = .051, p

> .05; Target Level of Control: F(1,139) = 1.085, p > .05; Referrals: F(1,120) = 3.324, p

> .05). The more experienced respondents did see three of the four groups of decisions as

more important: recommendations (4.25 versus 4.00), target level of control (4.30 versus

4.17), and referrals (4.22 versus 3.91). Medications decisions are the only group for

which the more experienced respondents indicated lessor importance (4.08) than less

experienced respondents (4.11).

Clinical Decisions Influenced by Psychosocial Factors and Physician Specialty (Internal Medicine and Family Medicine)

Last, I tested differences between frequency that psychosocial factors influenced

the four groups of clinical decisions, and physician specialty. The sample for this test was

only the physician sub-sample (n=36). Specialty is my independent variable and each of

the four groups of decisions is my dependent variable. The primary care physicians were

comprised of two specialties: family medicine and internal medicine. I used the one-way

154

ANOVA to investigate potential differences. I found no statistically significant difference

between specialty and importance of psychosocial factors for any of the four groups

(Recommendations: F(1,28) = .641, p > .05; Medications: F(1,31) = .300, p > .05; Target

Level of Control: F(1,26) = .952, p > .05; Referrals: F(1,19) = .897, p > .05).

5.8. Conclusion In this chapter, I described how practitioners use psychosocial information in the

course of making, or providing input into, type 2 diabetes care clinical decisions in the

outpatient setting. I introduced a cognitive map to describe the process of when and how

psychosocial factors influence clinical care decisions (see Figure 5.2). The cognitive map

is a visual representation of the key concepts of the cognitive process of psychosocial

information use. The four key concepts I described in the cognitive map are: 1)

considering clinical practice guidelines, 2) building and maintaining rapport with

patients, 3) triggers of psychosocial information consideration, 4) assessing the patient,

and 5) making the clinical decision.

Physicians are generally aware of the clinical practice guidelines but they readily

use clinical judgment to weigh the guidelines against the specific patient situation, which

is influenced by psychosocial factors. I describe how establishing an appropriate HbA1c

goal is the guideline most influenced by psychosocial factors. Physicians establish an

HbA1c goal above the guideline depending on financial strain, level of social support, or

patient preferences.

I described the substantial importance physicians place on establishing and

maintaining a patient-doctor relationship characterized by trust. This is necessary to grant

the physicians access to use psychosocial information to inform clinical decisions. I

155

described the methods they use to build the relationship, such as asking open-ended

questions, maintaining a safe environment without judgment, and empowering patients.

I outlined how psychosocial factors influence the specific four types of diabetes

care clinical decisions. Physicians use psychosocial information to inform decisions that

consider the patient’s circumstances which are driven by psychosocial factors such as

level of social support and financial strain.

I also showed the survey results on the influence of psychosocial factors on each

of the four types of clinical decisions, which included tests for association. I only found

statistically significant differences between practitioner role and frequency of influence

for making recommendations decisions. Nurse practitioners and diabetes educators

indicated that psychosocial factors influenced making recommendations more so than the

physicians indicated influence. I also examined the relationship between clinical

decisions, years of experience, and physician specialty. I did not find statistically

significant difference between years of experience and importance of psychosocial

factors for any of the four types of clinical decisions. Last, I found no statistically

significant difference between physician specialty and importance of psychosocial factors

for any of the four types of clinical decisions.

In the next chapter, I describe barriers to use of psychosocial information. I

describe how practitioners use EHR tools to document and retrieve psychosocial

information. I detail their confidence in using the EHR tools for psychosocial information

use, and frequency that these tools support the documentation and retrieval of

psychosocial information.

156

CHAPTER 6

BARRIERS AND FACILITATORS TO USE OF PSYCHOSOCIAL

INFORMATION

Discrete fields in an EHR fail to capture critical clinical information which would be captured in narrative….over-reliance on structured data capture risks information loss,

degrading value of the clinical record. Structured data capture can be at odds with the expressivity, workflow, and usability

factors preferred by clinicians. EHR systems often do not meet the needs of users, supporting downstream reporting

requirements at the expense of “clinically useful information”. ― HIMSS Health Story Project, November 2013

6.1. Introduction In this chapter, I describe barriers and facilitators to use of psychosocial

information. I touch upon other sources of psychosocial information, however I focus on

practitioners’ experience with electronic health record (EHR) tools to document and

retrieve psychosocial information in the course of providing diabetes care, specifically in

making, or providing input into, type 2 diabetes (T2DM) care decisions in the outpatient

setting. In this chapter, I answer my sixth research question:

RQ6: What are the barriers and facilitators to acquiring and using psychosocial information? How effectively do current tools (templates, data fields, free text) support the storage and retrieval of psychosocial information?

In answering this research question, I further develop the conceptual model of

psychosocial information access, initially depicted in chapter four (see Figure 4.3). I

157

include the additional psychosocial information physicians expressed was needed to help

inform clinical decisions, and ideas physicians offered to improve the documentation and

potential use of psychosocial information.

6.2. Facilitators to Use As described previously, the patient is a primary source of psychosocial

information. Physicians indicated that the quality of the relationship grants them access to

sensitive psychosocial information. I examined physician perceptions of the level of

accuracy of the psychosocial information that they received from patients. Physicians

also use the medical record to access and use psychosocial information to inform care

decisions. For instance, EHR alerts can prompt a referral to support services, and

physicians document psychosocial information that can serve as a reminder to them in

subsequent visits concerning a patient’s circumstances.

6.2.1. EHR Can Facilitate Use Physicians use the EHR to access psychosocial information to help inform clinical

care decisions. EHR systems incorporate alerts regarding psychosocial information that

can trigger a referral decision and in turn, help facilitate communication across the care

team. One internal medicine physician (P07) states how the EHR provides her with intake

information for new patients and alerts, which she can use to refer a patient to support

services; “all intakes get questions about fear of becoming homeless … that’s an

automatic flag in our system … we will refer them to Social Work to get them early

intervention … it’s a great system.”

Additionally, notes from referrals to other types of practitioners can bring

attention to psychosocial issues within the consultation. As P07 said, “If Mental Health

158

sees them … I make it a point to read the mental health note … to make sure there’s

nothing in there that I need to know about because I know it will affect their medical

care.”

The social history portion of the EHR may also trigger consideration of

psychosocial issues. A family medicine physician (P01) describes how psychosocial

information documented in social history helps physicians understand a patient’s

situation; “part of a doc’s social history [is] trying to figure out who’s at home. What

sort of contacts there are? What impacts those contacts are making on people? So for me,

it’s important to know where they’re working, who they’re living with, who’s important

to them.… It is [captured in the social history].”

Finally, information contained in free text notes may trigger consideration of

psychosocial issues. A family medicine physician (P14) describes how she uses the EHR

to document portions of the patient’s circumstances, the patient’s “story,” that serve as

prompts in future consultations. This information is documented during the consultation,

in pithy phrases. These help remind physicians of what may have been discussed during

the clinical visit; “it’s paraphrased because they’ll be telling me a story … about

depression, the drinking … their relationships, their home environment … and I don’t

need to put the whole story in … I just need a little tickler to remember. I look at it and

I’ll be like ‘Ah! Yes, I remember!’ It just comes flowing back, just the entire visit just

comes back … I will never remember if I don’t type it right then.”

Physicians may rely on their memory if retrieving psychosocial information in the

EHR is difficult, or if the needed information may not be there. As this physician said, “A

lot of things aren’t in the chart and certainly, there’s a place where you can put notes.

159

And you can put a note like, ‘Patient’s son recently committed suicide.’ And you can put

that in there kind of like in the family section. But I know I don’t routinely check that

section. I don’t think other people do either” (P15, Family Medicine). In such cases,

memory embedded within a long relationship with a patient may be sufficient from a

physicians’ perspective. Additionally, this information can be augmented at times with

confirmation from the patient during a visit; “I know my patients really well, so most of

that [psychosocial information] I’m actually … relying what’s on my head … if I vaguely

remember something, I’ll ask again … to verify what I remember, or clarify” (P08,

Internal Medicine).

6.2.2. Survey Results For the psychosocial information, respondents indicated as “Very Important” or

“Important” in making, or providing input into, clinical decisions, I asked them to

indicate all sources for this information. Respondents could indicate more than one

source. I gave respondents six choices: 1) patient, 2) family/caregiver, 3) other providers,

4) EHR, 5) other, and 6) no reliable source.

Survey respondents are most confident in the accuracy of psychosocial

information accessed from the patient, via consultation or interview. As shown in Table

6.2.2, respondents are next most confident in the accuracy of information from the other

providers or members of the care team. Next in confidence in accuracy is family and/or

caregivers. The patient, via screening tools and/or forms is equal to EHR. I isolate

physician responses in the table. None of the differences in responses among the three

sources is statistically significant.

160

Table 6.2.2 – Confidence in Accuracy of Psychosocial Information

Total (n = 159)

Primary Care Physicians

(n = 34)

Nurse Practitioners & Diabetes Educators (n = 125) p value

Patient, via consultation or interview 4.16 (.635) 4.15 (.558) 4.17 (.657) .865

Other Providers or Members of the Care Team

4.09 (.688) 3.91 (.712) 4.14 (.676) .092

Family / Caregivers 4.00 (.563) 4.06 (.600) 3.98 (.553) .493 Patient, via screening Tool or Forms 3.86 (.651) 3.85 (.500) 3.86 (.688) .930

EHR 3.86 (.757)a 3.82 (.717) 3.87 (.771)b .759 Note. Respondents were asked to indicate the frequency that they had confidence in the accuracy of psychosocial information they accessed from these sources. Responses were captured in a Likert scale: 5 – Always, 4 – Often, 3 – Sometimes, 2 – Rarely, 1 – Never. Standard deviations listed in parenthesis. a n=156. b n = 122.

6.3. Barriers to Use Although the EHR can facilitate use of psychosocial information, physicians

encounter considerable barriers in their efforts to utilize the EHR to use psychosocial

information for clinical decisions. Barriers include practice constraints (i.e., time

necessary to document and retrieve psychosocial information), and concerns about

consistency across the care team.

6.3.1. Clinical Practice Constraints Interview participants described circumstances that present barriers to using

psychosocial information to help inform diabetes clinical care decisions. I grouped them

according to clinical practice constraints and additional psychosocial information desired.

Physicians describe how the very nature of clinical practice in the primary care setting

presents barriers to access psychosocial information. Time constraints limit their ability

161

to probe psychosocial information. Also, learning about psychosocial barriers after

patients’ self-care has been negatively affected by them is an impediment to use.

Time Constraints Given that the patient is the most frequent source of psychosocial information, it

is perhaps unsurprising that interview participants consistently express the view that time

constraints are a major factor in their ability to access pertinent psychosocial information;

“in my opinion, that’s the biggest problem, that we as a primary care physicians never

have that much time, to … discuss [psychosocial information] with the patient” (P17,

Family Medicine). Although physicians feel they have the skill to probe on psychosocial

issues, time is a barrier to getting this information from the patient; “The reality of it […]

is the providers do not have the time. A lot of … question[s] [investigating barriers to

self-care] could go on forever” (P02, Internal Medicine). A family medicine physician

(P10) shares how she does not have the time to probe to assess health literacy in order to

explore potential barriers to self-care based on the patient’s comprehension of the

diabetes regimen; “I think I’m always feeling pressed for time. Of course that’s like a

complaint across the board. It’s like, ‘Where do you fit that in?’” A family medicine

physician (P06) expresses how time constraints restrict her ability to discuss pertinent

psychosocial information that could offer her important insights on barriers to self-care;

“lack of time … that’s an issue. There’s just less time to talk about things that maybe, I

could say, are specifically related to at least diet, exercise, and medication intake … they

are related to barriers for diabetes care.” Another family medicine physician (P10)

expresses frustration with not having the time to fully understand the patient’s

circumstances and must prioritize how she spends her consultation time; “Get[ing] …

162

people’s back story … is great, but it’s hard to do that and stay focused on your visit. So I

don’t always tend to [get the back story] … I do focus more on the medical piece … even

though … it’s important to focus on the back story of where the patient’s coming from,

but it’s hard to do all that.”

Physicians share specifics of the clinical consultation and how much time they

have to actually discuss potential psychosocial issues with their patients. One family

medicine physician (P06) discusses how she must allocate the limited time of the

consultation; “generally there are too many other problems to talk about. And there’s

limited time, just a 15-minute visit, which is like 5 minutes of discussion, 5 minutes of

exam. And then the last 5 minutes are like getting in, getting out, and then going to lab.”

An internal medicine physician (P03) points out that diabetes care is only a portion of a

primary care clinical visit, which is already time constrained, even with extra time

commonly allocated for a new patient visit; “if it’s a new patient, you get 10 minutes

face-time, right? If it’s … a[n] established patient, 5 minutes.… So, 10 minutes … for

addressing everything, including health maintenance … it’s really too short.… So

diabetes gets what, 30 seconds?”

Some physicians share how they are not incentivized to allocate their limited time

during the clinical consultation to probe and capture psychosocial information. Incentive

structures influence how they spend time with patients, and what they focus on “I’m not

rewarded based on the detail of the note but … that it meets certain coding requirements

… [and] that they’re done on time” (P08, Internal Medicine). An internal medicine

physician (P13) shares his perspectives on how performance measures may influence

how he allocates time with patients:

163

Many of the current metrics [incent physicians] to spend more time with a person who has a (HbA1c) 9.2 than the person who has a like the 8.7 or 8.5 that really we should be trying to aggressively get to 7.… maybe I really should spend more time with that person who has the A1c of 8.5 … tapping into some of the psychosocial stuff … to figure this out and really help them and understand what the barriers are and help them get to where they need to be.

6.3.2. Concerns about Consistency Across the Care Team Physicians express concerns about variation between care team members in how

they document psychosocial information. Uncertainty surrounding how psychosocial

information is documented causes physicians to speculate if their documentation is useful

to other members of the care team. A family medicine physician (P15) questions if other

physicians follow the same procedure she does in documenting psychosocial information

she deems important; “you can write … the psychosocial things that you think would be

helpful for somebody to know …’Patient like has a severe history of abuse’… I think that

kind of thing, we will put in the problem list … I don’t know if everybody does, but I will

put it there.”

Because of this uncertainty and variation in use, physicians question the value of

taking time to search the record for pertinent psychosocial information, as what they are

seeking may not be there. Rather than checking the record, they choose to get information

from the patient; “it’s easier to just ask the patient [about ability to pay for medications],

than it is to dig through someone else’s notes that may or may not have documented that”

(P08, Internal Medicine). Additionally, physicians note that psychosocial information that

they document in the record may be overlooked by other practitioners who provide care

to the same patient. For example, an internal medicine physician (P08) expressed

frustration with important psychosocial information she documented in the record which

164

was missed by her partners at her clinic and those who provided care to the patient during

a hospital stay:

Sometimes [I’ll] type in some comments … under the unstructured problem list or under past medical history … or in the social history. But by and large, a lot of that will get lost…. I know that from experience because if my patients go to the hospital or get seen by my partners in clinic, there might’ve been something really important that I knew about them and their psychosocial situation and that I documented using this pretext that didn’t get picked up by my partners when they were taking care of the patient … that’s probably the biggest thing, there’s always the risk that what I documented didn’t click with them [i.e., partners on care team] … what I meant in the social history.

6.3.3. Questions Regarding Accuracy of Psychosocial Information Accessed I examined the perceived accuracy of information because it is an important

dimension of access. All study participants indicate fairly high levels of trust in the

accuracy of psychosocial information that they access as they make, or influence, clinical

decisions for diabetes patients. On an unprompted basis, physician interviewees

expressed the concern that, at times, they may question the accuracy of the psychosocial

information that patients provide. An internal medicine physician (P02) with twenty years

of experience, explains his scrutiny of psychosocial-related information this way, “I’ll be

honest with you … most of the time, a lot of patients, I have a feeling, I can’t validate it,

but my own intuition is, sometimes they tell us what we wanna hear…. [when] they

disclose to me that they are compliant and have all … “yes’s” to what I’m asking. That’s

when I start to be dubious at something … they’re not being upfront.” Physicians may

also evaluate the information they receive from clinical tests against their experience with

some patients who attempt to skew results of common diabetes clinical measures, like

fasting blood sugar (FBS). A family medicine physician (P12) with over thirty years of

experience shares, “people study for tests as I always say. So they know they’re gonna

165

get their fasting blood sugar checked, they’ll starve themselves for 3 days and say, ‘My

sugar’s fine.’ And then they’re so starved they’ll go out and their sugar will be 500, and

they’ll be in comatose by the time they get home.”

There are risks associated with accessing social support information from family

and caregivers. Gleaning level of support can be delicate if the family member(s) are

present during the clinical visit. An internal medicine physician (P11) describes situations

in which it may be difficult to get information:

Family support is not very easy to find out. That’s the toughest ... I’ve seen patients … The son is here [points to the left], the mother is here [points to the right] … You can literally see them arguing right in front of you. If [they] say, “Okay, my mom doesn’t support me,” and the mom is like next to [them] ... I think, that’s one thing which is difficult to get out. It’s difficult to get out of a patient … family support.

Health Literacy - Problems with Access Physicians describe barriers to accessing accurate information about health

literacy. Low health literacy can influence the accuracy of the information the patient is

asked to provide; “I think both in terms of health literacy and just literacy, in general …

there’s a lot of paperwork that patients … fill out … I think that is … a definite gap, both

in terms of my own screening, and just in terms of things that we’re not doing as well as

we could do” (P15, Family Medicine). Another family medicine physician (P16) explains

that lack of access is because disclosing low literacy may be too personal, compared to

other personal information patients are comfortable sharing; “I have a fair number of

conversations with people about domestic violence, about rape, about things like that …

weird sexual practices, multiple sexual partners … that people are very open about …

I’ve never had a conversation like that about literacy.”

166

Health Literacy - Problems with Assessment Although level of health literacy is an important psychosocial factor, it is difficult

to assess. A family medicine physician (P10) describes challenges to assessing literacy;

“we don’t always do a really good job about finding out who can’t read and write … it is

kind of hidden, you have to kind of tease it apart … you end up giving them handouts and

things to write on … it kind of miss[es] the mark … we don’t really assess patients’

literacy really well…. We don’t have in a formal way of assessing it. So that’s [a]

limitation.” This assessment gap was echoed by other physicians, “[we] don't have a

good handle on health literacy, how much they understand” (P02, Internal Medicine).

Another family medicine physician (P15) revealed that her assessment of a patient’s

health literacy may be higher than it actually is, “I'm probably not as good... I mean ... I

wish I were better about screening for [health literacy] … particularly, because when

you read things about health literacy, it's often a lot lower than I think it is.”

Physicians describe various techniques they use to attempt to assess health

literacy level, which they associate with level of education and more general literacy.

One family medicine physician (P06) describes his approach:

I [have] never used a screening tool for health literacy, there are some that are available, but I would generally gauge it by patients.... I would have them look at the bottles with me and I'd say, ‘So this bottle says taking twice a day… two tablets twice a day’… [assessing] literacy for me was more around their bottles and what can they tell me [about] the information on the bottle.

A family medicine physician (P10) describes the complexity of trying to assess this

important psychosocial factor:

A lot of times, it is kind of hidden [level of health literacy] … you have to kind of tease it apart … and we don't always do a really good job about finding out who can't read and write… You end up giving them handouts and things to write on, and it [can] kind of miss the mark.” She continues with doubts about her care

167

team’s ability to assess health literacy, “we don't really assess patient's literacy really well… I don't know … we don't have a formal way of assessing it. So that's a limitation … we're not necessarily always doing a great job with the health literacy.” (P10, Family Medicine)

6.3.4. EHR Presents Barriers to Use Physicians share various barriers they experience to using specific EHR tools (i.e.,

data fields, templates, and free text) to document and retrieve psychosocial information. I

arrange these EHR barriers in two areas: 1) time required to document psychosocial

information and 2) the design of current tools.

Time Required to Document Psychosocial Information Psychosocial information may be discussed at various times over the course of the

clinical consultation, but it may not be documented in the medical record due to the time

it takes to enter psychosocial information. Therefore, physicians may not use the EHR

tools to document it. A family medicine physician (P12) discusses experiences across

different care teams; “Not everybody, myself included, would always include a very

robust social history in writing. It just took too much time.” An internal medicine

physician (P08) cites time as a barrier to documenting more extensive psychosocial

information on the patient’s “story”, which includes pertinent information such as

barriers to following the recommended diabetes regimen, “Because of time, sometimes

typing really fast it might be something like, ‘DM discussed difficulties with lifestyle,’ and

that may be all I say, and I may know a 10 minute story about that but I didn’t put it in

[the EHR].” Physicians describe having to choose between spending time with their

patients and documenting psychosocial information in the record. They choose to spend

their limited time talking to their patients, rather than documenting information in the

168

record; “it takes a lot of time … it’s so time-consuming [to capture psychosocial

information in free text] that I’d rather talk to my patient for the vast majority of the time

and then spend a minimal time writing about what we spoke about, and then go on to my

next patient where I spend time talking to them” (P15, Family Medicine).

Concerns about How the Information will be Used Physicians also may not document self-care behavior because of concerns that

their notes may be interpreted inaccurately by other practitioners or may result in

stigmatization of patients. Therefore, as this physician explains, they may be reluctant to

document this information in the EHR:

[a patient] may … have a history of noncompliance but I don’t usually … write the reason in the chart. I don’t know…. We tend not to put … in the chart … things that might implicate the patient. (P15, Family Medicine)

Design of EHR Tools Participants also highlight the belief that the design of specific EHR tools — data

fields, templates and free text — presents barriers to both documenting and retrieving

psychosocial information. Some psychosocial information physicians express influences

clinical care decisions, such as level of social support and financial strain, is not in the

record. Physicians do not believe these tools facilitate easy documentation, or use, of

relevant psychosocial information.

Documentation EHR tools do not enable easy documentation of psychosocial information at the

specificity required. The tools can be too complex and layered. A family medicine

physician (P16) describes the difficulty in documenting rationale for a target goal for

HbA1c, which can be driven by psychosocial factors not documented in the record:

169

I don’t think it’s as explicit as we’d like it to be … ideal would be … some place in the chart where we said this patient’s A1c goal is 8 or 9 and … clarifying … the reason that we’re not doing tight control.… It may be at the end of a note, but it’s not some place that’s easy to find.… I’m not even sure … that I do it the way that I think would be ideal.

Current tools do not enable physicians to document specific psychosocial information

pertinent for diabetes care decisions. An internal medicine physician (P08) describes the

limitation of data fields:

The structured fields that we have are things like marital status and education level, which are important, but it’s actually not as important, because health literacy and literacy are 2 different things that can’t be predicted, unfortunately, by college education…. [Also] there [is not] … a way to specifically document what their goals of treatments are, what their self-perceived barriers are.

A family medicine physician (P06) describes the limitations on what types of

psychosocial information the data fields can document. For example, level of social

support is not captured. This is highly pertinent psychosocial information for clinical

decisions:

There is no tab for social support in the medical record … I see the diabetes hypertension, hypercholesterolemia, back pain, abdominal pain. I see symptoms, and I see diagnoses, but I don’t see psychosocial issues … there are tabs for notes and problem list and medications and orders, but there’s no tab for psychosocial issues.

An internal medicine physician (P09) discusses further limitations in data fields for

documenting specific psychosocial information; “you can code for … narcotic addiction

or tobacco, nicotine addiction, but there’s not one for poverty or economic, food

insecurity or shelter insecurity. So we missed the boat.”

Templates are also not designed to capture the specific psychosocial information

that may be pertinent to making clinical decisions. The structured format of templates do

not lend themselves to capturing information concerning the patient’s background and

170

living situation. A family medicine physician (P16) describes how the templates do not

enable practitioners to capture the patient’s story sufficiently:

Epic [EHR] has a big section for social stuff. But it’s not all that usable. I don’t quite know why it doesn’t work, but it doesn’t work … nobody uses it … [Because] psychosocial history is very story oriented, it’s not very checkbox oriented.

Psychosocial information documented in free text fields may be difficult to use, because

it is not documented in phrases that provide insight, or is difficult to locate within the

text. This is especially the case when other members of the care team may have

documented it; “somebody else might find that it’s [i.e., the psychosocial information]

cryptic or it’s very broad. It doesn’t actually give the granular detail or it’s simply not in

a place they can find [it] … and if they did, it’s a note, it’s a phrase” (P08, Internal

Medicine).

Retrieval/Use Physicians express considerable challenges to retrieving and using psychosocial

information based on capabilities of EHR tools, specifically that which is contained in

data fields and templates. A family medicine physician (P15) describes difficulty in

retrieving psychosocial information concerning mental health status; “there’s no place in

MyChart [EHR] where I can find out what somebody’s PHQ-9 was (Patient Health

Questionnaire - a diagnostic tool for mental health).” To address the limitations of these

current tools, physicians simply depend on other means to use psychosocial information.

Since the current capabilities of data fields are inadequate, physicians may rely on their

memory to retrieve and then use psychosocial information; “with some of my patients …

[I’m] worried about nutrition … Unfortunately there’s no great way to actually capture

that in the EMR so a lot of it goes into unstructured fields in the HPI (History of Present

171

Illness). Or it [was] buried 3 years ago … [in] a progress note, that’s not very helpful …

it’s more in my head than it is in an accessible way in the chart” (P08, Internal

Medicine).

Templates are difficult to use for retrieving specific psychosocial information.

They are not designed for use with the broad range of relevant psychosocial information

physicians may deem relevant to inform clinical decisions. In addition, their design is

characterized by several layers to access. A family medicine physician (P12) describes

how difficult it can be to retrieve relevant psychosocial information; “Sometimes it [isn’t]

all that easy to find where … [psychosocial information] … is captured … If their

daughter’s involved in their care or ... who’s the primary caretaker?... The way they are

designed … complex … the templates [are] oftentimes … buried under multiple layers.”

6.3.5. Availability of Psychosocial Information to Support Clinical Decisions In this section, I outline the availability of psychosocial information. I describe

how it can be difficult to evaluate financial barriers to self-care. I describe survey

findings concerning the frequency that respondents have the information they need.

Financial barriers Financial barriers can be difficult for physicians to ascertain. Simply knowing

income level or payer status may not be sufficient to inform a clinical decision. Ability to

pay for medications is a more precise financial barrier that physicians deem important.

An internal medicine physician (P07) with over a decade of experience based in

Michigan, states, “I don’t necessarily [use] income as much as ... the patient’s ability to

pay for their medications.”

172

Availability of Psychosocial Information Although respondents indicate that they frequently use psychosocial information

for clinical care decisions, it is frequently unavailable. As shown in Table 6.3.5

psychosocial information is most frequently available for medications decisions.

Respondents indicated 3.7 out of a 5 point Likert scale (5 – Always, 4 – Often, 3 –

Sometimes, 2 – Rarely, 1 – Never). Also, 59.0% of responses indicated “Always” or

“Often”. Making recommendations decisions are next (3.69/5; 58.9%). Other decisions

are next (3.67/5; 67.3%). Target level of control decisions are next (3.63/5; 54.8%),

followed by making referrals (3.62/5; 54.9%). I isolate physician responses in the table.

None of the differences in responses are statistically significant.

Table 6.3.5 – Frequency with which Respondents Have the Psychosocial Information They Need

Total (n = 168)

Primary Care Physicians

(n = 39)

Nurse Practitioners & Diabetes Educators (n = 129) p value

Medications 3.70 (.765)a 3.82 (.756) 3.66 (.767)b .267 Making Recommendations 3.69 (.700) 3.79 (.801) 3.66 (.667) .289

Other Decisions 3.67 (.923)c 3.50 (1.000)d 3.73 (.905)e .464 Target Level of Control 3.63 (.734)f 3.74 (.910) 3.59 (.671)g .335 Making Referrals 3.62 (.704)h 3.72 (.759) 3.58 (.686)i .301

Note. Respondents were asked to indicate how often they had the psychosocial information they need when making, or providing input into, clinical decisions. Responses were captured in a Likert scale: 5 – Always, 4 – Often, 3 – Sometimes, 2 – Rarely, 1 – Never. Standard deviations listed in parenthesis. a n=161. b n = 122. c n = 52. d n = 40. e n = 122. f n = 166. g n = 127. h n = 164. i n = 125.

Timing of Access to Psychosocial Information Physicians describe when psychosocial information is disclosed is also a barrier to

incorporating psychosocial factors into clinical decisions. Participants describe how they

may find out after a period of time that a patient is not taking medications as directed,

173

because they cannot afford them. A family medicine physician (P10) describes cases

when patients disclose financial barriers during follow-up visits; “I’ve had patients that

didn’t have money to buy their medicines.… It frustrates me because they’ll be out of

their medicines for 2 months … [they] tell me … at their follow up visit. I [say], ‘Could

you just call us? We will help you. Let us know instead of not taking your insulin for 2

months.’” Another family medicine physician (P15) describes how she may find out over

time that a patient does not understand aspects of the care regimen:

Sometimes if things are poorly controlled … [I’ll] have them repeat … back to

me … [to] make sure that they … understand what I’m saying.… I will go through their

medications 1 by 1 and find out they’re not taking them … or not the way it was

prescribed…. I’ll find out that way, but that’s after the fact. It’d be nice if I knew that up

front.

6.4. Final Conceptual Model of Psychosocial Information Access To summarize results of psychosocial information access, I built upon the initial

model shown in chapter four (see Figure 4.3) to create the final conceptual model of

psychosocial information access (see Figure 6.4). I have added clinical decisions, and

availability and accuracy of psychosocial information.

174

Top PFs Source of PI How Accessed

1. Financial Strain (4.84)*

3. Life Stressors(4.57)

5. Social Support(4.53)

2. Mental Health(4.62)

4. Food Security(4.55)

Patient(43.0%)**

Family/Caregiver (28.4%)

Other Providers(15.5%)

EHR(11.6%)

• Data fields: mental health, payor status

• Prompting

• Engaging others• Listening• Questioning /

Clarifying

• Questioning (open-ended)

• Listening

• Asking/Calling (nurse, pharmacist)

6. Health Literacy(4.53)

Practitioner Role (Physicians vs. Other)

Decisions Influenced

Medications(4.09)

Recommendations (4.18)

Target Control Levels(4.26)*

Referrals(4.13)

PI – Psychosocial InformationPFs – Psychosocial Factors* - Mean of responses to Likert scale responses: 5 – Very Important, 4 – Important, 3 – Neither Important nor Unimportant, 2 – Unimportant, 1 – Very Unimportant** - % of Total Source Responses. Respondent could indicate more than one source.*** - Responses captured in a Likert scale: 5 – Always, 4 – Often, 3 – Sometimes, 2 – Rarely, 1 – Never

Confident in Accuracy of PI1. Patient – consultation (4.16; 89.3%)***2. Other providers (4.09; 81.8%)3. Family/caregivers (4.00; 84.3%)4. Patient-self-reported tools (3.86; 76.1%)5. EHR (3.86; 69.9%)

Availability of PI

1. Medications (3.70; 59.0%)***2. Making Recommendations (3.69; 58.9%)3. Target level of control (3.63; 54.8%)4. Making Referrals (3.62; 54.9%)

Access to Psychosocial Information

Figure 6.4: Final Conceptual Model of Psychosocial Information Access

6.5. Additional Psychosocial Information Desired Physicians discuss how they would like to have access to specific psychosocial

information because information they currently can access is not comprehensive.

Moreover, participants contend that having access to more detailed psychosocial

information would better inform clinical care decisions. I next outline the additional

psychosocial information desired, but currently not accessed, according to the four

groups of psychosocial factors.

Sociodemographic Psychosocial Factors Physicians express interest in learning more about sociodemographic psychosocial

information. Financial barriers, dietary practices based on cultural norms, level of

175

education, and religion are among the additional information desired to help inform

clinical decisions because these psychosocial factors may limit patients’ ability follow

recommended self-care behavior. An internal medicine physician (P07) expresses the

belief that the current practice of documenting occupational classification (i.e. exempt or

non-exempt) is not a sufficient indicator of financial stress; “understanding their

financial situation might help. Because non-exempt (status) might not help. It might not

give you that granular information … It won’t.” Further, documenting payor status does

not provide adequate insight on potential barriers to self-care due to financial strain. A

family medicine physician (P15) says, “I don’t know anything about their income …

unless … they’re Medicaid … sometimes … that’s kind of [an] indication.” Another

family medicine physician (P15) describes various psychosocial information she would

like beyond what is currently documented. Additional information concerning education

level, language, religion, transportation, and level of social support would also help

inform clinical care; “highest grade of school … their preferred language … preferred

religion … How do they get to their clinic appointments? Who is their contact person?

Sometimes who they’ll give you as their contact person may be totally different from the

person who I think is the person who is key … they may give their husband’s name, but

he’s as clueless as the patient.” Another internal medicine physician (P11) details

additional psychosocial information desired:

I would like to know more about the resources they have, first of all. Second, their mental activity levels. And third … family support … how much does your family support … maintenance of your disease.... Your lifestyle … religious beliefs … we had a lot of cultures back home [India] which have a different eating pattern … people do a lot of fastings everyday.

176

An internal medicine physician (P08) expresses need to know more about dietary

practices based on culture, this would help her level of understanding of the patient’s

circumstance; “What’s their diet like? Why is their diet like that? Is it because of

financial restrictions? Is it because they’re Japanese and me telling them not to eat white

rice is just not a practical thing to tell them?”

Social Relationships/Living Condition Psychosocial Factors Physicians describe additional information desired concerning a patient’s home

environment. One internal medicine physician (P08) explains the additional information

desired, “more of a description about their home situation or competing priorities” (P08,

Internal Medicine). An enhanced understanding of the home environment can help

physicians understand dietary and smoking habits, or potential exposure to second hand

smoke; “if they’re diabetic I’m even more concerned if they’re smoking or … if there’s

people in the house that are smoking … Their living situation is really important … If

you’re living with other people or eating with the other people, you’re eating as well…. if

there’s kids at home … there’s juice at home, cookies, cakes, all that stuff” (P07, Internal

Medicine). P08 also states how additional information on how diabetes may be impacting

the patient’s family would be helpful:

If you could get patients to tell you something about honestly how supported they feel ...”Does this chronic disease influence your family?”… And, “In what ways does it help and in what ways it does not help your relationship with your family members?” [that] is the evidence you really need …. to make some decisions … because the diabetes not only creates a medical problem, it creates a social problem. (P08, Internal Medicine)

Neighborhood/Community Participants spoke to how knowledge of community factors can help distinguish

barriers to care from the patient’s desire to follow the self-care regimen. An internal

177

medicine physician (P07) shares how additional information about the patient’s perceived

safety of their neighborhood may help her understand specific barriers which helps her

provide support in addressing them; “their ability to exercise, either the safety of the

neighborhood versus their personal willingness to be able to engage in exercise to help

control their diabetes.” Another internal medicine physician (P08) describes the

difficulty with determining potential barriers to self-care due to neighborhood

psychosocial factors. Knowing more detailed information about community safety, such

as crime statistics for specific areas of a community (e.g., census tract) would help her

assess potential barriers. Simply knowing the patient’s address may not be sufficient:

Knowing the communities … having some sense of the broader context of things … that’s really fraught with problems, right? Because then you’re totally relying on stereotypes …. we know that zip code … does correlate pretty well with socioeconomics … [but]sometimes it plays against you…One of my residents bought a house not too far from Frandor (Michigan) … I’m from this area (central Michigan)…. I made the assumption that she might be in an unsafe neighborhood.… It turns out … it’s a lovely neighborhood.… I had made some assumptions that were completely incorrect…. On the other hand … they have a child care center that’s right across from the hospital, and there was a shooting at the houses directly across the street from [it] … So, you just never know.

6.6. Improvement Ideas In light of current barriers, physicians offer specific ideas to improve the

capabilities of current EHR tools to better support their use of psychosocial information

to inform clinical care decisions. Enhancing detail on current documentation would be

helpful to understand cultural influence on self-care practices. One family medicine

physician (P12) discusses how “ethnicity”, a common field in the record, is not a precise

enough indicator of cultural norms that could inform diabetes care clinical decisions,

especially for patients who are not born in the United States:

178

we label Hispanics, Latinos as being this [one ]group … but it’s enormously diverse and very, very, very, very, very, different … the Nepalese, the Iraqis … their foods are different.… their diet … how … they eat … their preferences, their cultural beliefs … how they used to access care in their country, what their expectations may be.

He suggests including country of origin and ethnicity in the record, “you could capture

… at registration … what their ethnic group was and where they were born….”

The family medicine physician (P12) suggests how to make this information more

easily accessible. Linking ethnicity to detailed background information would give the

practitioner a more comprehensive understanding of culture, and health beliefs, “[country

of origin and ethnicity] could then tie to some very specific information regarding health

that a clinician could either read prior to the visit, [or] after the visit, but would be linked

to that particular chart.” Making this information easily usable in the record would help

inform clinical decisions. An internal medicine physician (P08) suggests “pop-up” boxes

that appear, for example, under the problem list; “it would be interesting if … on a

problem list … for diabetes … [you could] tie [in] some comments.… If you hover over

… a whole list of things [appear] … the psychosocial information … ’During the winter

doesn’t exercise because no safe way to do so’… things like that. [If I see that] I’m … not

going to worry so much about glycemic control.”

6.7. Survey Findings – EHR Tools Insights from the physician interviews concerning the perceived barriers to

documenting and retrieving psychosocial information helped inform the online survey,

which I administered to a larger, more diverse group of practitioners. I used the online

survey to further probe perceptions concerning capabilities to document and retrieve

psychosocial information for three specific EHR tools: templates, data fields, and free

179

text. I asked respondents to indicate how frequently these tools support the

documentation and retrieval of psychosocial information. Also, I asked respondents to

indicate the frequency with which they felt confident using these tools to document and

retrieve psychosocial information to make, or provide input into, care decisions. I report

results for all survey responses; I also isolate physician survey responses to examine any

statistical significance between physician respondents and the combined group of nurse

practitioners and diabetes educators.

6.7.1. Frequency with which EHR Tools Support Documentation and Retrieval Respondents most frequently believe that free text tools support the

documentation of psychosocial information, as shown in Table 6.7.1.1. I isolate physician

responses in the table. The differences in responses for data fields are the only for which

the differences are statistically significant.

Table 6.7.1.1 – Frequency that EHR Tools Support Documentation of Psychosocial Information

Total (n = 154)

Primary Care Physicians

(n = 34)

Nurse Practitioners & Diabetes Educators (n = 120) p value

Free Text 3.84 (.864) 3.91 (.793) 3.83 (.886) .607 Data Fields 3.57 (.783) 3.85 (.892) 3.49 (.733) .017 Templates 3.51 (.842) 3.74 (.898) 3.45 (.818) .081

Note. Respondents were asked to indicate the frequency that each of the three EHR tools support the documentation of psychosocial information. Responses were captured in a Likert scale: 5 – Always, 4 – Often, 3 – Sometimes, 2 – Rarely, 1 – Never. Standard deviations listed in parenthesis.

Respondents most frequently believe that free text tools also support the retrieval

of psychosocial information, as shown in Table 6.7.1.2. I isolate physician responses in

the table. There are no statistically significant differences in responses.

180

Table 6.7.1.2 – Frequency that EHR Tools Support Retrieval of Psychosocial Information

Total (n = 152)

Primary Care Physicians

(n = 34)

Nurse Practitioners & Diabetes Educators (n = 118) p value

Free Text 3.68 (.918) 3.68 (.843) 3.68 (.942) .993 Data Fields 3.61 (.887)a 3.67 (.777)b 3.59 (.917) .675 Templates 3.59 (.930) 3.50 (.992) 3.62 (.914) .514

Note. Respondents were asked to indicate the frequency that each of the three EHR tools support the retrieval of psychosocial information. Responses were captured in a Likert scale: 5 – Always, 4 – Often, 3 – Sometimes, 2 – Rarely, 1 – Never. Standard deviations listed in parenthesis. a n=151. b n = 33.

6.7.2. Confidence in Using EHR Tools Respondents most frequently indicate that they are confident that free text tools

support the documentation of psychosocial information, as shown in Table 6.7.2.1. I

isolate physician responses in the table. There are no statistically significant differences

in responses.

Table 6.7.2.1 – Confidence in Using EHR Tools to Support Documentation of Psychosocial Information

Total (n = 153)

Primary Care Physicians

(n = 34)

Nurse Practitioners & Diabetes Educators (n = 120) p value

Free Text 4.16 (.852) 3.97 (.810)a 4.21 (.859) .155 Data Fields 3.77 (.892) 3.76 (.855) 3.77 (.906)b .962 Templates 3.76 (.911) 3.62 (.954) 3.80 (.898)b .309

Note. Respondents were asked to indicate the frequency in which they had confidence in using the three EHR tools to support the documentation of psychosocial information Responses were captured in a Likert scale: 5 – Always, 4 – Often, 3 – Sometimes, 2 – Rarely, 1 – Never. Standard deviations listed in parenthesis. a n=33. b n = 119.

Respondents also most frequently indicate that they are confident that free text

tools support the retrieval of psychosocial information, as shown in Table 6.7.2.2. I

181

isolate physician responses in the table. There are no statistically significant differences

in responses.

Table 6.7.2.2 – Confidence in Using EHR Tools to Support Retrieval of Psychosocial Information

Total (n = 152)

Primary Care Physicians

(n = 34)

Nurse Practitioners & Diabetes Educators (n = 118) p value

Free Text 3.80 (.914) 3.65 (.684) 3.85 (.948) .261 Data Fields 3.74 (.895) 3.68 (.684) 3.76 (.949) .557 Templates 3.72 (.931) 3.53 (.896) 3.77 (.937) .183

Note. Respondents were asked to indicate the frequency in which they had confidence in using the three EHR tools to support the retrieval of psychosocial information. Responses were captured in a Likert scale: 5 – Always, 4 – Often, 3 – Sometimes, 2 – Rarely, 1 – Never. Standard deviations listed in parenthesis.

6.8. Conclusion In this chapter, I answered my sixth research question by describing how

practitioners experience barriers to the use of psychosocial information. I focused on the

effectiveness of current EHR tools in documenting and retrieving psychosocial

information in the course of making, or providing input into, type 2 diabetes care clinical

decisions in the outpatient setting. Using the online survey data, I quantified the

frequency that these tools support psychosocial information documentation and use, and

the confidence level practitioners indicate in using these tools to support documentation

and use.

I described how physicians used EHR tools to help inform clinical decisions. My

analysis of the interview data enabled me to understand then describe how physicians use

current EHR tools to help inform referral decisions and facilitate communication of

clinical information across the care team. I also described how physicians can use their

notes in the EHR to help trigger their memory in recalling specific psychosocial

information concerning a patient case.

182

Next, I described concerns physicians expressed regarding how clinical practice

constraints present barriers to use. I described how quality concerns, time constraints,

timing of access to psychosocial information, and design of EHR tools present barriers.

Time pressures limit their ability to discuss psychosocial information with patients. I

described the tradeoff physicians make between documenting psychosocial information

and spending time with their patients. Time constraints can impede their ability to use the

EHR to document and retrieve psychosocial information. Practice incentives can also

present barriers to use. Physicians are not incentivized to capture the patients’ situation

pertaining to psychosocial factors.

I also described how inconsistency across the care team presents barriers.

Physicians question the benefit of taking the time to search the record for psychosocial

information that may have been documented by another team member, given that this

information is not documented consistently. Consequently, they may ask the patient

instead of using the EHR. Physicians also expressed concerns with documenting

psychosocial information in the record because other members of the care team may not

interpret their notes accurately. For example, documenting the reasons for a patient’s

unhealthy self-care behavior may stigmatize the patient.

I outlined the specific barriers physicians described to using the EHR for

documentation, then retrieval of psychosocial information. I arranged these barriers in

two areas: 1) time required to capture psychosocial information and 2) the design of

current EHR tools: data fields, templates, and free text. Physicians express that

documenting the patient’s social history is time consuming. As a result, they do not

document some psychosocial information. The design of EHR tools is a barrier to both

183

documenting and retrieving psychosocial information. For instance, concerns about

access to healthy foods or rationale for a certain HbA1c goal may include pertinent

psychosocial information, but current tools are not designed to document or retrieve this

information easily.

I then described the additional psychosocial information that physicians desire,

but cannot access and subsequently use. Physicians indicated that this information is

either unavailable or only becomes available too late, after the time when they could have

made adjustments to help address barriers to care. For example, patients may not initially

disclose a psychosocial barrier (e.g., financial strain that prevents them from taking

medications as directed), which may impede timely response to addressing these

problems. I also described how physicians express desire to know more about

psychosocial factors such as: cultural norms that may affect dietary habits, health literacy,

their home environment responsibilities, and occupational situation.

I described specific ideas physicians shared to improve the documentation and

retrieval of psychosocial information. It would be helpful if tools could accommodate a

more granular level of psychosocial information (e.g., “ethnicity” may not be sufficient to

capture cultural beliefs or dietary preferences).

Last, I used my survey data analysis to describe the use and perceptions

concerning specific EHR tools. I described how respondents indicated that free text tools

most frequently support the documentation and retrieval of psychosocial information,

when compared to data fields and templates. I concluded with describing confidence in

using these tools to document and retrieve psychosocial information. Again, respondents

184

most frequently indicated high confidence free text tools for documentation and retrieval

of psychosocial information.

In the next chapter, I summarize the study. I discuss my conclusions and offer

implications for future work. I also outline the strengths and limitations of the

investigation.

185

CHAPTER 7

DISCUSSION AND IMPLICATIONS

The pioneers have begun their work. It is far from finished. New fields, new enterprises

are visible. The times call for the high spirit of the courageous pioneers among physicians, scientists, and nurses.

― Lillian Wald (1867 – 1940) Founder of the Visiting Nurse Service in New York City

Investments in larger systems of economic, environmental, and social support produce

health and support individuals’ quest for well-being. ― Elizabeth H. Bradley

7.1. SUMMARY OF FINDINGS

Psychosocial Factors, and their relevance (RQ1) Practitioners believe that psychosocial factors are relevant to consider in making

clinical decisions because they affect self-care behavior known to influence diabetes

outcomes. Physicians emphasized how the complexity of the diabetes care regimen

requires that the patient is able to effectively “manage their lives.” They strongly consider

the following six psychosocial factors because they present barriers to self-care, and can

make it difficult for the patients to follow the recommended care regimen: financial

strain, mental health status, life stressors, food security, social support, and health

literacy. Financial strain is of particular note as it influences medication self-care. Also, a

patient’s mental health status and level of social support are of import, since these factors

may drive establishing higher HbA1c targets than are recommended in practice

186

guidelines, in part to avoid the potential risk of harm associated with hypoglycemic

episodes.

My findings for practitioner perceptions of psychosocial factors are consistent

with the literature I detailed in chapter two (section 2.3.1). For example, financial strain

presents barriers to access medications, and influences medication self-care behavior

(Jotkowitz et al., 2006; Kangovi et al., 2013).

How practitioners access psychosocial information (RQ2) The patient is the most frequent source of psychosocial information. Patients

provide physicians information to help them understand potential barriers to self-care,

such as financial issues which prevent them from following recommended medication

self-care. In addition, physicians describe how their patients state that their other

responsibilities can cause them to focus on activities they determine are more important

(i.e., occupational responsibilities) than diabetes self-care. Physicians also rely on the

patient’s family and caregiver(s) to access pertinent psychosocial information such as

mental health status and level of social support. Other members of the care team provide

support in assessing the nature of the particular barrier to self-care (i.e., health literacy),

and offer support in making clinical decisions to address it. The EHR is also used as a

source of psychosocial information. It can serve as a direct information source (i.e.,

regarding mental health status) and can also help provide information (i.e., using refill

information) that can trigger a conversation about potential barriers.

These interview findings are generalizable based on the survey participants

indicating the patient as the primary source (43%), followed by the family / caregivers

(28.4%), other providers (15.5%), and the EHR (11.6%). These findings are consistent

187

with the literature I cited in chapter two (section 2.4.2), outlining how the patient is a

source of psychosocial information (Kassirer et al., 2010). They also are also consistent

with the literature indicating that the EHR is a source of psychosocial information for

medication refill information (Doyle et al., 2012). These findings contribute to the

literature by linking how practitioners access specific psychosocial information to the

sources from which they acquire it, and how it helps inform their outpatient care

decisions.

Conceptual Model of Psychosocial Information Access The final conceptual model of psychosocial information access (see Figure 6.4)

provides a framework for further investigation of the associations between psychosocial

factors, information sources, and clinical decisions. The depiction of these associations

fills a current persistent gap in clinical decision making models discussed in chapter two

(section 2.4). The conceptual model describes how psychosocial information is accessed

for the four types of diabetes clinical decisions investigated.

How practitioners use psychosocial information (RQ3) Physicians use psychosocial information to help them develop a care regimen

appropriate for the patient situation. This information helps them: assess the patient,

understand the influencers on their self-care behavior, determine their capabilities (i.e.,

the degree to which they are able to “manage their lives”), and assess their level of

comprehension of required self-care practices. Physicians consider the diabetes clinical

practice guidelines (CPGs), but psychosocial factors can greatly influence deviations

from them, in particular for establishing an appropriate target for HbA1c. This interview

finding was confirmed in the survey results, which show that target level of control

decisions are very frequently influenced by psychosocial factors.

188

Cognitive Map of Psychosocial Information Use The cognitive map of psychosocial information use (see Figure 5.2) represents

how physicians incorporate psychosocial information into their clinical decisions. It

includes their thought processes to determine how CPGs may not directly apply to the

patient situation and depicts triggers for consideration of psychosocial factors. The map

also depicts how physicians access sensitive, pertinent psychosocial information from the

patient, which informs the clinical decisions.

Physicians access psychosocial information by developing rapport with patients,

facilitated by seeing them over a period of time. Continuity of care is important in

developing the relationship. Physicians express the importance of “getting to know” their

patients over multiple clinic visits; this helps them develop a trusting relationship with

their patients. Through this connection, physicians are able to access personal, sensitive

psychosocial information which patients share. Physicians detail the techniques they use

to build the rapport necessary to create and maintain these trusting relationships; they

provide a safe environment, free from judgment. They also empower patients with shared

decision making by acknowledging the critical role patients have in influencing clinical

decisions.

Psychosocial information informs practitioners’ assessment of the patient and all

four types of diabetes clinical decisions I examined. This information informs subsequent

judgments about the feasibility of treatment options, clinical risk, and patient needs. For

example, treating an at-risk patient who experiences financial strain influences an

assessment of patient capabilities and a related feasibility judgment, which then results in

a decision to establish an HbA1c target lower than the target recommended in the

guidelines. Further, physicians take financial strain and related capability and feasibility

189

judgments into account when making medications decisions, in particular selecting type

of insulin, which has implications for financial strain based upon out-of-pocket costs.

Physicians acknowledge that their decisions may be clinically suboptimal (i.e., selecting a

less expensive insulin), but they reason that these decisions are appropriate based on their

assessment of the patient situation.

When are Psychosocial Factors Considered (RQ4) Psychosocial factors are specifically considered in three distinct types of

circumstances: 1) chronic circumstances, which are based on ongoing situations (i.e., at-

risk patients, when a patient is not reaching clinical goals); 2) new circumstances, which

are based on an emergent situation (i.e., new patient, initial diagnosis); and 3) a change in

circumstances, based on a variation in the patient’s situation (i.e., A1c “spike”, sudden,

unhealthy self-care practices). They are not particularly considered if the patient seems to

be doing well and meeting their clinical goals.

Practitioner characteristics’ influence on use of Psychosocial Information (RQ5) Physicians are consistent in their perceptions of the influence of psychosocial

factors. Across the three clinical specialties represented in my sample—and level of

experience from resident to thirty years of practice—all used psychosocial information to

help inform their clinical decisions. The final conceptual model of psychosocial

information use (see Figure 6.4) shows how the most pertinent psychosocial information

is accessed, for which decisions, and barriers to use.

My survey findings extend the literature cited in chapter two (section 2.6)

regarding practitioner role and influence of psychosocial factors. I found statistically

significant difference between physicians and other practitioners. Nurses tend to believe

190

more frequently than physicians and diabetes educators that psychosocial factors

influence self-care practices (Funnell, 2006; Rasch & Cogdill, 1999). Also, I found

statistical significant differences in practitioner role for influence for specific clinical

decisions. Nurse practitioners consider psychosocial factors more frequently than

physicians for medications decisions. Diabetes educators consider them more frequently

than physicians for referral decisions. These findings extend the knowledge concerning

practitioner consideration of psychosocial factors to diabetes educators, beyond

physicians and nurses. Furthermore, I found no statistical significance in examining

difference of influence of psychosocial factors on clinical decisions between physician

responses by specialty, nor for years of practice experience.

Barriers to Use – Efficacy of EHR Tools (RQ6) Current EHR tools can present barriers to use. I have segmented barriers in two

areas: 1) time required to document psychosocial information, and 2) the design of

current tools.

Physicians express that time constrains present barriers to psychosocial

information use. Since the patient is the most frequent source of psychosocial

information, it follows that interview participants express that time restrictions hinder

access to psychosocial information relevant to their clinical decisions. Physicians are

quite confident in their skill to access this information, yet they generally do not have the

time to probe on psychosocial information which could provide useful context on self-

care behavior and potential barriers to care.

The design and capabilities of current EHR tools also present barriers to use. The

time required to document psychosocial information is of particular note. The tools are

191

complex, and require physicians to maneuver through several layers to document

pertinent information. Data fields and templates are not designed to capture psychosocial

information such as level and nature of social support, or barriers to healthy food options.

Physicians describe how spending the time required to document psychosocial

information in the EHR can take time away from the patient consultation. Thus, they

generally chose to talk to their patients.

7.2. Introduction In the previous three chapters, I introduced the initial conceptual model of

psychosocial information access (see Figure 4.3) which describes: the relevance of

specific psychosocial factors, the sources of psychosocial information, and for which

diabetes care clinical decisions this information is used. I also introduced the cognitive

map of psychosocial information use (see Figure 5.2), which depicts how practitioners

use this information to inform clinical decisions. In chapter six, I described barriers and

facilitators to acquiring and using psychosocial information, which include practitioner

perspectives on how current EHR tools may support, or hinder, their use of psychosocial

information. I also introduced the final conceptual model of psychosocial information

access (see Figure 6.4).

In this final chapter, I discuss and interpret my four major findings: 1)

psychosocial information is not considered when patients are stable and well controlled,

but it is considered when a patient: has persistent, poor glycemic control; is new to the

practitioner or has a new diabetes diagnosis; or has worsening of glycemic control, 2)

access to psychosocial information is granted through dialogue in the context of an

ongoing, trusted relationship, 3) awareness of psychosocial information may trigger

192

decisions to personalize HbA1c targets, pursue less aggressive treatment plans or

augment guideline-concordant treatment with actions to address barriers to care, and 4)

current EHR designs are not optimized for capturing and retrieving qualitative and

situationally dependent psychosocial information which tends to come in a narrative

form. I revisit the initial conceptual framework I introduced in chapter two (see Figure

2.2) and include the study results to help illustrate where these major findings address

gaps in the literature (see Figure 7.2). The dark blue boxes in the center of the framework

depict how my major findings concerning psychosocial information address current gaps

in the literature described in chapter two. The framework can be used to guide future

work to investigate the associations between health outcomes, and the three key

influencers of outcomes which my findings show are influenced by psychosocial factors:

use of CPGs, clinical decision making, and efforts to improve patients’ self-care

practices.

* - PFs – Psychosocial Factors * - PI – Psychosocial Information * - CPGs – Clinical Practice Guidelines ** - Clinical Decision Marking mediates influence of PFs on self-care (i.e., when recommending resources for food, housing, etc.)

Figure 7.2: Final Conceptual Framework of Psychosocial Factors in Outpatient Diabetes Care

193

Taken together, my four major findings help further the understanding of when

and how psychosocial information is used by practitioners as they provide diabetes care

in the outpatient setting. I extend the insights from previous studies which describe: the

influence of various psychosocial factors on diabetes outcomes (Arigo, Smyth, Haggerty,

& Raggio, 2014; Bielamowicz et al., 2013; Moulton, Pickup, & Ismail, 2015; Piette &

Kerr, 2006; Safford et al., 2005; Selby, 2010), information use for clinical decision

making in outpatient care (Del Fiol et al., 2014; Doyle et al., 2012), and CPG use (Evert

et al., 2014; Kirpitch & Maryniuk, 2011). I describe how practitioners use psychosocial

information, which includes their rationale for diverging from practice guidelines in

attempts to lessen the influence of barriers to recommended self-care due to psychosocial

factors. For each major finding I also examine the implications. I conclude with a

discussion of the strengths and limitations of the study.

7.3. Four Major Findings

1. Psychosocial Information is not considered when patients are stable and well controlled, but it is considered when a patient: has persistent, poor glycemic control; is new to the provider, or has a new diagnosis; or has worsening of glycemic control

If the patient is stable and well controlled, psychosocial information may not be

particularly relevant; however, in three specific situations psychosocial information is

considered. In these circumstances, practitioners use psychosocial information to help

them understand potential barriers to self-care and to inform their clinical decisions to

help patients address them. First, psychosocial factors are relevant under chronic

circumstances, such as when a patient experiences persistent, poor glycemic control or

194

when treating at-risk patients. Second, psychosocial information is considered under new

circumstances, such as when seeing a new patient or when a patient is newly diagnosed

with diabetes. Third, practitioners consider psychosocial information when there is a

change in circumstances, such as when there is a worsening of glycemic control indicated

by an increase (“spike”) in HbA1c and/or when there is sudden, unhealthy self-care

behavior.

The literature I cite in chapter two (section 2.5.1) describes that physicians may

consider the applicability of guidelines according to patient cases—a patient’s gender and

age influenced frequency of foot exams (McKinlay et al., 2013). However, my findings

revealed that these stable patient characteristics do not exclusively trigger consideration

of psychosocial factors, and this study makes a novel contribution in that it is the only

study of which I am aware that describes the specific, shifting clinical circumstances

(e.g., a patient’s worsening glycemic control and/or self-care practices) in which

practitioners consider psychosocial information to inform diabetes care decisions in the

outpatient setting.

While I purposively sampled practitioners who are sensitive to psychosocial

factors based on their experience with at-risk patients, this finding is of particular note

given that patients’ lives are dynamic rather than static, and that their financial

circumstances, social support, stressors and mental health may change over time. Further,

it is useful to understand the specific circumstances when providing psychosocial

information will be helpful to practitioners given that they are already exposed to large

amounts of clinical data designed to support clinical decisions. The sheer volume of

clinical data can exacerbate existing time demands (Haase, Follmann, Skipka, &

195

Kirchner, 2007), which practitioners can find overwhelming (A. Hall & Walton, 2004;

Hübner-Bloder et al., 2011; Wright, McCoy, Henkin, Kale, & Sittig, 2013). This suggests

the value of just-in-time information that is provided precisely during those clinical

circumstances in which providers in my study deemed important.

Knowing when psychosocial information is used by practitioners to guide their

care decisions is an important contribution to the objective of providing practitioners with

the information they need, when they need it; this is critical since doing so remains a

persistent challenge (Bass, DeVoge, Waggoner-Fountain, & Borowitz, 2012). Future

research should focus on making pertinent psychosocial information available during

these three circumstances which we now understand trigger consideration of psychosocial

factors.

Further, these findings extend previous clinical decision making literature,

specifically clinical decision support, through an enhanced understanding of the

circumstances in which psychosocial information is used. They connect the influence of

psychosocial factors to clinical decision making and help explain how clinical goals may

be established outside of targets specified in the guidelines.

2. Access to psychosocial information is granted thorough dialogue in the context of an ongoing trusting relationship

In chapter five (section 5.4.1), I describe how physicians develop these

relationships by using specific methods to demonstrate caring and establish and maintain

a safe environment, void of pejorative language. They maintain the patient’s privacy and

empower them by acknowledging their role in the care regimen, which encourages shared

decision making. These findings are consistent with prior research I described in chapter

196

two (section 2.4.1) which describes how the quality of the patient-doctor relationship

increases the likelihood of disclosure of psychosocial information (Higgs, 2008;

Robinson & Roter, 1999a, 1999b). Chronic disease patients with multiple providers—

resulting from discontinuous care, multiple chronic conditions, or frequent

hospitalizations—are at greater risk of low continuity of care (Haggerty et al., 2012;

Kripalani et al., 2007). Continuity of care is an important contributor to a patient’s ability

to follow self-care recommendations (Bodenheimer, 2008). Therefore, trusting

relationships are an important determinant of regimen adherence, which influences

diabetes outcomes (Delamater, 2006; Ritholz et al., 2014). Moreover, effective

communication between the T2DM patient and their practitioner results in increased

practitioner knowledge of patients, a positive predictor of healthy self-care practices and

of chronic disease outcomes (Morrison et al., 2013; Singh-Manoux, 2003; Weyrauch,

Rhodes, Psaty, & Grubb, 1995).

These findings provide additional support for the value of measuring and

evaluating the quality of the practitioner-patient relationship, which has been shown to

have a statistically significant effect on healthcare outcomes in studies focused on cross-

cultural differences (De Faoite & Hanson, 2015). Findings also provide additional

information as to why physicians value the quality of the practitioner-patient relationship;

this may be because it grants them access to psychosocial information that enables them

to access information pertinent to their clinical decisions which they believe result in

better outcomes. There is a need for additional research examining how the quality of the

practitioner-patient relationship influences chronic disease outcomes associated with

chronic conditions (J. M. Kelley, Kraft-Todd, Schapira, Kossowsky, & Riess, 2014).

197

These findings contribute to the current literature by suggesting a new mechanism

through which strong communication and a trusting, nurturing relationship potentially

influences outcomes. The quality of the relationship enables access to pertinent

psychosocial information, which subsequently is used to inform personalized care

decisions intended to lessen the influence of barriers to recommended self-care caused by

psychosocial factors. A stable, trusting relationship ensures the consistent availability of

pertinent psychosocial information necessary to inform these personalized care decisions,

which study participants believe result in better outcomes.

Physicians clearly described specific techniques they find effective to developing

the relationships that enable access to sensitive psychosocial information. Insights

concerning how physicians build and maintain trusting patient-doctor relationships

should be used to further the understanding of mechanisms by which these relationships

are established and maintained. These insights can be used to help extend research on the

efficacy of training resources concerning patient engagement, and measuring its impact

on outcomes.

3. Awareness of Psychosocial Information may trigger decisions to personalize HbA1c targets, pursue less aggressive treatment plans or augment guideline-concordant treatment with actions to address barriers to care

Practitioners use psychosocial information to help them understand and address

barriers to self-care behavior, either through personalization of treatment plans or actions

to address patients’ support needs. Accordingly, psychosocial factors help inform the

clinical decisions practitioners believe are most appropriate for the specific patient

circumstance. I now discuss two particularly notable clinical decisions informed by

198

psychosocial information derived from the qualitative analysis: 1) determining the

applicability of CPGs, and 2) medications decisions.

Applicability of Clinical Practice Guidelines (CPGs)

Practitioners use psychosocial information to determine the applicability of

clinical practice guidelines (CPGs) based upon their understanding of the patient’s

current self-care behavior. A patient’s financial circumstance, mental health status, and

low level of social support may drive the decision to establish higher HbA1c targets than

specified in the guidelines, in part to avoid the potential risk of harm associated with

hypoglycemic episodes. They may also pursue less aggressive treatment plans or

augment guideline-concordant treatment with actions to address barriers to care. These

include referrals to prescription assistance, food support, counseling, and transportation

aid.

These findings are important to consider in order to precisely evaluate clinician

performance against clinical practice goals. They implore us to consider a more

expansive perspective on what constitutes quality care. Indeed, findings imply that

evaluating clinical performance solely against HbA1c targets is shortsighted. Research

investigating approaches to report diabetes quality do not consider practitioners clinical

decisions based upon attempts to address psychosocial barriers to self-care (Halladay et

al., 2014; O'Connor et al., 2011; Oxendine, Meyer, Reid, Adams, & Sabol, 2014).

Ignoring the diversity of patient circumstances, and practitioner appreciation of them,

could result in performance indicators that miss recognition of practitioners who are

particularly attuned to their patients’ situations, and are making clinical decisions

accordingly.

199

Medications decisions

I found that physicians’ medications decisions in particular can be considerably

influenced by psychosocial factors. Physicians attempt to reduce the influence of barriers

to recommended medication self-care due to financial strain, low social support, or

mental health issues. These decisions include selecting type of insulin, based on patient

out-of-pocket cost and insurance coverage. Physicians acknowledge that such decisions

may be clinically suboptimal, but they contend that they are appropriate based on their

assessment of the patient situation, which can be informed by their understanding of the

influence of psychosocial factors.

These findings are consistent with the literature concerning how financial strain

presents barriers to access medications, and influences medication self-care behavior

(Jotkowitz et al., 2006; Kangovi et al., 2013). However, these findings extend the current

breadth of the clinical decision making literature in describing that, in certain patient

circumstances, physicians choose to make “suboptimal” clinical decisions in their efforts

to lessen the influence of barriers to self-care stemming from psychosocial factors; they

do so intentionally. Furthermore, findings contribute to the clinical decision making

literature by describing how consideration of psychosocial factors may affect the rate of

practitioner adherence to clinical practice guidelines (Appiah et al., 2013; Gaucher,

Lantos, & Payot, 2013; Goldstein, Lavori, Coleman, Advani, & Hoffman, 2005).

Confidence in decisions to help address self-care barriers

Study participants are confident in making clinical decisions to help address

barriers to recommended self-care, if they are aware of barriers the patient may be

experiencing. They regret learning “after the fact” about barriers to self-care caused by

200

psychosocial factors, rather than feeling incapable to support patients who may have

difficulty with “managing their lives” due to psychosocial barriers.

In chapter two (section 2.6), I cite research which describes that although

physicians may recognize the critical influence that psychosocial factors have on self-

care behavior and health outcomes, they do not necessarily feel equipped to address

psychosocial needs (Institute of Medicine, 2008; Robert Wood Johnson Foundation,

2011). In the context of diabetes care in particular, practitioners may not be confident in

their ability to address the psychosocial barriers encountered by some diabetic patients.

Again, my sampling approach offers a possible explanation for the divergence. My

findings stem from a sensitized group of physicians, nurses and diabetes educators who

may recognize the importance of psychosocial factors more so than practitioners who

participated in these previous studies. The practitioners in my study sample have

considerable experience in care settings acquainted with at-risk patients who present with

psychosocial barriers to self-care, thus these settings may have established resources they

call upon to help address barriers.

The literature describes how practitioners may make clinical decisions in order to

avoid contextual errors, which can occur when a patient’s situation is not accurately

assessed (Weiner et al., 2010). My findings provide insight on circumstances when

practitioners may not have an accurate understanding of psychosocial barriers. Although

study participants expressed confidence in knowledge of psychosocial barriers related to

neighborhood setting, they may not be correct in their assumptions about barriers due to

psychosocial factors. For example, in chapter six (section 6.5), I described how an

internal medicine physician shared her misconceptions concerning crime in two areas of

201

Michigan; both areas in which she was familiar. Also, in chapter six (section 6.3.3), I

described how physicians have real concerns about their ability to accurately assess

health literacy. Extending this work to understand the results of these actions could

support clinical decisions support systems (CDSS) enhancement. For example, providing

specific crime statistics at a more granular level (i.e., census tract) might enable

practitioners to make more informed clinical decisions concerning neighborhood safety, a

known psychosocial factor that influences self-care behavior (Cadzow et al., 2014; C. R.

Clark et al., 2013; Jack et al., 2012; Krishnan et al., 2010).

Future research should also evaluate practitioner beliefs about the influence of

psychosocial factors on self-care practices against what at-risk patients experience.

Recently published research has described the effort that at-risk patients make in their

attempts to adhere to the chronic care regimen; this “invisible work” performed by

patients is done outside of the healthcare system. Therefore, this effort may not be

recognized by practitioners (Senteio & Veinot, 2014).

Results can help support practitioners in offering strategies to address self-care

barriers that actually influence diabetes outcomes. As capabilities are developed to

improve the collection and use of psychosocial information, understanding the efficacy of

resulting clinical decisions in addressing barriers will continue to be important. These

results should help guide practice efforts to improve the documentation and use of

psychosocial information that practitioners deem is important. Moreover, capabilities

should determine the influence of specific psychosocial factors on self-care based on

extant literature. This would help to confirm the practitioners’ perceptions of the

relevance of psychosocial factors and their influence on self-care.

202

Since in this study I purposively sampled practitioners attuned to psychosocial

factors, in subsequent research, it would be informative to study less-sensitized

practitioners who are not necessarily actively engaged in care with large proportions of

at-risk patients. Practitioners who serve patients who are not at-risk may perceive the

relevance of psychosocial factors quite differently. Understanding how these practitioners

may consider psychosocial factors as they make clinical care decisions can inform how

these findings can be incorporated in a broader set of circumstances, including for clinical

decisions beyond diabetes care. Also, interviews with non-physician practitioners who

make or influence medications decisions would enable further understanding of potential

associations of a broader set of clinical decisions based on practitioner role, again for

diabetes clinical decisions and those for other chronic conditions.

Further study concerning tools and training to standardize the collection and use

of pertinent psychosocial information across the entire care team could help increase the

understanding of the influence of psychosocial factors. In particular, these tools should

include screening for health literacy, which have been developed for diabetes care but are

not widely used (Schillinger, Grumbach, Piette, Wang, Osmond, Daher, Palacios, et al.,

2002). Making these tools available enables using the EHR for secondary research, such

as in the field of clinical epidemiology. Such efforts could also be used for investigations

of how access to psychosocial information influences its use, and the impact of its use as

measured by subsequent clinical decisions and outcomes. Results would provide a more

comprehensive perspective on psychosocial facilitators and barriers to self-care.

These results can also be used to help document and disseminate how

practitioners support patients who are experiencing barriers to care based on psychosocial

203

factors. Future efforts to address diabetes outcomes should use practitioners’ considerable

knowledge of the influence of psychosocial factors on self-care, and the steps taken to

address self-care barriers, to support the design of tools and processes to support clinical

decision-making, and inform interventions designed to provide support for patients who

may experience barriers to self-care. Further, they can be used to help train resident

physicians in how to support patients who they learn are experiencing barriers to care due

to psychosocial factors.

4. Current EHR designs are not optimized for capturing and retrieving qualitative and situationally dependent psychosocial information which tends to come in a narrative form

Study participants indicated that psychosocial factors frequently influence their

clinical decisions, but they frequently do not have the information they need to inform

these decisions. Barriers to access are primarily due to the current design of EHRs which

make it difficult to use psychosocial information. Of the sources I investigated,

participants indicated that the EHR is the least frequently used source of psychosocial

information. This is primarily due to barriers to documenting and retrieving psychosocial

information.

Findings from this research extend previous work by highlighting concerns

regarding the documentation of psychosocial information in the EHR, which results in

inconsistencies and concerns about data quality. Study participants indicate that a vital

challenge to psychosocial information use lies in not learning of psychosocial barriers

soon enough. This is related to the limitations of existing tools. In chapter six (section

6.3.5), I reported survey results for psychosocial information use for clinical decisions.

204

Respondents indicated they have the psychosocial information they need at a lower

frequency than psychosocial information is considered. Physicians stated that

psychosocial information is frequently not documented in, nor is it accessed from, the

electronic medical record due to time constraints and design of current tools. This is

consistent with research on the inpatient care setting I describe in chapter two (section

2.4.2), which also described that practitioners tend to not document psychosocial

information (Zhou et al., 2009). Also, while recent research has described how use of the

EHR improves the quality of clinical notes for documentation (Burke et al., 2015), my

work is focused on psychosocial information use. That study only contrasted quality of

EHR notes against handwritten notes.

These findings illuminate a potential gap in unmet needs. In chapter two (section

2.4.2), I cited literature concerning unmet information needs in primary care (Bass et al.,

2012; Del Fiol et al., 2014). If the majority of information needs are not met, it follows

that these decisions are not supported adequately with current, available tools. This

finding supports the building motivation to further develop EHR capabilities to identify

and capture psychosocial information, as evidenced by the 2014 Institute of Medicine

reports recommending that EHR tools should improve the documentation of psychosocial

information (Institute of Medicine, 2014a, 2014b). However, linking gaps in information

needs to types of decisions that are currently unsupported is a unique contribution.

An important consideration in addressing barriers to documentation and use of

sensitive psychosocial information is that doing so would potentially make it accessible

to other members of the care team. While physicians emphasized the importance of

protecting the patients’ privacy from other patient family members they may know, or

205

may treat, they did not express explicit reservations about other members of the care team

having access to sensitive psychosocial information told to them. They did however

express concerns about stigmatizing patients if reasons for self-care were documented in

the record. Providing access to sensitive psychosocial information across the care team

has the potential to breach trust. Further research should examine how patients perceive

the availability of their sensitive psychosocial information to other members of the care

team. Findings would inform best approaches to sharing this information in the outpatient

care setting, as a supplement to research previously mentioned focused on the inpatient

care setting (Zhou et al., 2009, 2010).

Last, future research should investigate approaches to provide specific

information which could quantify psychosocial “health”. A psychosocial “index” could

help address the current gap of practitioners lacking an instrument to assess and track a

patient’s psychosocial health. This instrument could serve much like clinical diabetes

measures currently in wide use, such as: HbA1c, fasting blood sugar, and mental health

measures like the Beck depression scale. The development of this psychosocial index

could also give patients an opportunity to provide information concerning their current

psychosocial health. The psychosocial factors I outline in Appendix B: Psychosocial

Factors could serve as a starting point for developing the index, which should include

specific psychosocial information elicited from the patient, the care team, and from other

information sources, such as information about the crime rates in a patient’s

neighborhood. Furthermore, I anticipate a platform for future research based on potential

implementation of the aforementioned Institute of Medicine recommendations to enhance

EHR capabilities concerning the use of psychosocial information (Institute of Medicine,

206

2014a, 2014b). Upon development of such an index, investigating its influence on

clinical decision making and subsequent health outcomes should be examined. Health

outcomes of interest include: HbA1c, cost of care, patient satisfaction, patient

engagement, and practitioner satisfaction.

7.4. Strengths and Limitations

Strengths A strength of this study is in the sequential exploratory design. In general, mixed

methods research is recognized for its capacity to leverage both qualitative and

quantitative methods (Curry et al., 2009). This design was well suited to investigate the

degree to which insights developed from a few individuals can be generalized to a larger

sample. This approach helped me address each of my research questions as I explored a

considerable gap in understanding in an area of clinical decision-making—the role of

psychosocial factors. It is well established in the literature that psychosocial factors

influence self-care practices associated with diabetes outcomes; however, there was no

common description for how they may influence care decisions. This design was

appropriate given the gap I sought to fill.

Another strength of this study is that it enables a comprehensive exploration of

the factors that influence clinical decisions heard directly from the physicians as they

describe how they make diabetes care decisions as part of their regular clinical

responsibilities. Including a variety of practitioner roles and levels of experience in the

sample positioned the study well to describe how clinical decisions are made. Also,

geographic diversity of practice settings enabled examination across clinical

environments, as well as across various at-risk patient populations. In addition, my

207

recruitment of a sensitized group of practitioners facilitated in-depth investigation of the

phenomenon of interest.

The generalizability of findings to a larger population was confirmed for findings

regarding the specific psychosocial factors of import, the decisions influenced, and the

situations in which psychosocial information is considered. This increased confidence in

the interview findings concerning the relevance of specific psychosocial factors, and their

influence on specific clinical decisions.

Limitations The main limitation of this study is that it relies on a sensitized group of

practitioners who are experienced with diabetes care for at-risk patients. Hence my

sample was fairly homogeneous in their perspective in the importance and use of

psychosocial factors. My findings should thus be interpreted considering that practitioner

participants are familiar with psychosocial barriers to care; therefore, participants may

have more experience and confidence in their ability to address them. In addition, care

teams experienced with providing care for at-risk patients may have established clinical

processes in place to help practitioners address psychosocial barriers, such as processes

and relationships in place to refer patients to community-based resources. These may not

be present in other clinical settings. Consequently, findings are not generalizable to all

practitioners, particularly those without experience with at-risk patients.

Another limitation of this study design is that it relies on what the practitioner

says influences their clinical care decisions, which is subject to bias associated with

desirability and recall. Also, what individuals indicate influences their actions may differ

from their actual beliefs and actions; this is an established phenomenon in qualitative

208

research (LaPiere, 1934; LeCompte & Goetz, 1982; Liebow, 1967). The findings are also

limited because they are not considered against data contained in patients’ medical

records. Examining the clinical record to understand what clinical decisions are actually

made would help investigate specific clinical decisions on a large scale, such as referral

practices and medication decisions.

A further limitation of the initial qualitative data collection and analysis step is

that I only interviewed physicians; I did not interview patients. Understanding the

patients’ perspectives on associations between psychosocial factors and self-care would

help to understand what and how psychosocial factors may facilitate, or present barriers

to, self-care practices. Further, I did not interview other members of the care team which

could shed light upon how psychosocial factors may be considered in making, or

influencing, diabetes care clinical decisions. There are two reasons for this. First, since

physicians frequently have control over the greatest number of clinical decisions, a focus

on these practitioners appeared warranted. I sought to understand physicians’ perceptions

of psychosocial factors, and describe them in the context of the “sensitizing concepts”

drawn from the literature (Blumer, 1969). Second, I was seeking to fill the considerable

gap in the literature as to how psychosocial factors may influence specific diabetes care

clinical decisions, emphasizing patient populations who may be subject to barriers to care

due to psychosocial factors. I focused on physicians with experience with these patient

populations in order to get their perspectives based experience with numerous patient

cases.

I used an online survey for a major portion for my data collection. There are

several advantages and disadvantages to using surveys in research. Advantages include

209

breadth of coverage in a relatively short amount of time (K. Kelley et al., 2003). The

survey sample, drawn from three different professional organizations, provided a high

degree of exposure to practitioner perspectives. This wide scope enabled me to analyze a

more representative sample, making it more generalizable to a population of primary care

physicians, nurse practitioners, and diabetes educators, all experienced in diabetes care.

However, the low response rates (29.8%) meant that the sample might differ

systematically from the populations from which these samples were drawn. Given the

topic of the study, I again anticipate that the respondents may have been a group that was

particularly sensitized to psychosocial issues.

7.5. Conclusion In this investigation, I detailed how practitioners access and use psychosocial

information to inform diabetes care clinical decisions. Study participants were well

acquainted with the influence of psychosocial factors, and used psychosocial information

during new or problematic circumstances. Psychosocial information was granted

primarily through dialogue that emerged in a trusting and ongoing clinical relationship.

Participants generally used psychosocial information to address self-care barriers through

personalization of HbA1c targets, less aggressive treatment plans and linking patients to

supplemental resources. However, they were not always aware of relevant psychosocial

factors. Findings revealed opportunities for improving the capture and retrieval of

psychosocial information; such improvements would necessarily focus on better

accommodation of information in narrative form. As developers continue to evolve

electronic tools in their capabilities to collect, analyze, and provide personal health

information, psychosocial information should also be included because of its direct

210

influence on clinical decision-making, and potentially ultimately on diabetes-related

health outcomes.

211

APPENDICES

212

Appendix A Definitions

Outpatient care – where the patients stay at the hospital or clinic from registration to

discharge over the course of a single day. It is the most common setting where diabetic patients access their health care (Willens et al., 2011)

Electronic Health Record (EHR) – digitally stored healthcare information throughout

an individual’s lifetime with the purpose of supporting continuity of care, education, and research. The EHRs may include such things as observations, laboratory tests, medical images, treatments, therapies; drugs administered, patient identifying information, legal permissions (Ajami & Arab-Chadegani, 2013; Hillestad et al., 2005)

Hypoglycemia – low blood sugar, if not treated can lead to seizure, unconsciousness Patient-Centered Decision Making (PCDM) - incorporates clinically relevant, patient-

specific circumstances and behaviors, that is, the patient’s context, into formulating a contextually appropriate plan of care. The objective was to develop a method for analyzing physician-patient interactions to ascertain whether decision-making is patient-centered. (Weiner et al., 2014)

Primary Care – The Institute of Medicine (IOM) defines primary care as ‘the provision

of integrated, accessible health care services by clinicians who are accountable for addressing a large majority of personal health needs, developing a sustained partnership with patients, and practicing in the context of family and community (Donaldson, Yordy, Lohr, & Vanselow, 1996). Specialties associated with primary care include: family practice, general practice, general internal medicine, and pediatrics (Gorman & Helfand, 1995)

213

Appendix B

Psychosocial Factors

This is a description of the psychosocial factors investigated in the practitioner survey.

Sociodemographic psychosocial factors

1. Financial strain lack of resources that impact access to food, safe housing, transportation, or medications

2. Employment job demands that may influence self-care (e.g. work hours, type of job, ability to take time off for self-care)

3. Payor status/Type of insurance

if a patient has insurance coverage, type of coverage, private insurance, government health insurance (Medicare, Medicaid, Military, State-specific plans, Indian Health Service) (United States Census Bureau, 2014)

4. Culture and spirituality cultural norms and traditions which include dietary practices, faith beliefs, and practices (Kittler, 1995; Kittler & Sucher, 1995a)

5. Other responsibilities work, family responsibilities (e.g. serving as caregiver for elderly adults), self-care activities

6. Level of education level of formal schooling 7. Literacy Literacy is the ability to use printed and written

information to function in society, to achieve one’s goals, and to develop one’s knowledge and potential (National Center for Education Statistics, 2014)

8. Country of origin U.S. born or non-U.S. born, Immigrant status 9. Level of English

proficiency ability to understand and speak English

Psychological psychosocial factors

1. Mental health status appearance; manner and approach; orientation, alertness, and thought processes; mood and affect (National Institute of Mental Health, 2014)

2. Life Stressors negative events, chronic strains, traumas (Thoits, 2010)

214

3. T2DM perceptions-beliefs

perceptions of the relative quality-of-life effects of complications and treatments (Huang, Brown, Ewigman, Foley, & Meltzer, 2007)

4. Health Literacy measure of patients’ ability to read, comprehend, and act on medical instructions (Schillinger, Grumbach, Piette, Wang, Osmond, Daher, Palacios, et al., 2002)

5. Health Numeracy the degree to which individuals can obtain, process, and understand the basic quantitative health information and services they need to make appropriate health decisions (Institute of Medicine of the National Academies, 2014)

Social Relationship/Living Conditions psychosocial factors

1. Social Support social isolation, social connections, support includes four dimensions: appraisal support, informational support, instrumental support and emotional support. (Funnell, 2010)

2. Threat of violence - from abusive relationship(s)

a pattern of coercion, physical abuse, sexual abuse, or threat of violence in personal relationships (Krug, 2002)

3. Threat of violence - from community

perception of violence caused by violence-inducing or violence-protecting conditions (Sampson, Morenoff, & Raudenbush, 2005)

Neighborhood/Community psychosocial factors

1. Patient’s Rural/Urban/Suburban residence setting

residency setting, may influence health care utilization or access (Spoont et al., 2011)

2. Neighborhood residence

physical aspects of a neighborhood that influence a patient’s ability to purchase products (food), enable mobility, and interact and informally monitor another’s behavior (D. A. Cohen et al., 2003)

3. Housing security Stable housing, access to affordable, safe housing (Johns Hopkins Center to Eliminate Cardiovascular Health Disparities, 2014a)

4. Food security access to fresh, healthy, and affordable food (Walker, Keane, & Burke, 2010)

5. Access to transportation

barriers which can lead to missed appointments, missed or delayed medication use (Syed, Gerber, & Sharp, 2013)

6. Access to places to exercise

accessible facilities, include perception of safe places to exercise (Bennett et al., 2007)

215

Appendix C

Type 2 Diabetes Clinical Decisions

This is a description of the type 2 diabetes care decisions investigated in the provider survey. Target level of control care decisions

1. Establish target goal for blood glucose 2. Establish target goal for HbA1c 3. Incorporate input from patient in setting goal 4. Other Target Goal Decision(s)

Medications care decisions

1. Select a specific medication (e.g. long acting insulin or short acting insulin) 2. Select a brand, or a generic medication 3. Start a patient on first oral diabetes medication 4. Adjust oral diabetes medication dosage 5. Add an additional (2nd, 3rd, 4th, etc.) oral diabetes medication 6. Start a patient on injectable insulin 7. Adjust injectable insulin dosage 8. Start a patient on non-insulin injectable diabetes medication (e.g. GLP-1) 9. Adjust non-insulin injectable diabetes medication (e.g. GLP-1) 10. Reduce complexity of medication regimen 11. Other Medication Decision(s)

Making Referrals care decisions

1. Refer a patient to specialty care (e.g. endocrine, mental health, etc.) 2. Refer a patient to support services outside the organization (e.g. social work,

housing, transportation, etc.) 3. Refer a patient to support services within the organization (e.g. social work,

housing, transportation, etc.) 4. Refer a patient to diabetes education 5. Refer a patient to a dietitian and/or nutritional information 6. Other Referral Decision(s)

Making Recommendations care decisions

1. Make dietary recommendations 2. Make physical activity recommendations 3. Recommend a patient’s caregivers understand what is required of patient 4. Frequency of clinical visits 5. Other Recommendation Decision(s)

216

Appendix D

Outline of ADA Standards of Care s

I. CLASSIFICATION AND DIAGNOSIS A. Classification B. Diagnosis of diabetes C. Categories of increased risk for diabetes (prediabetes)

II. TESTING FOR DIABETES IN ASYMPTOMATIC PATIENTS A. Testing for type 2 diabetes and risk of future diabetes in adults B. Screening for type 2 diabetes in children C. Screening for type 1 diabetes

III. DETECTION AND DIAGNOSIS OF GDM - Gestational diabetes mellitus IV. PREVENTION/DELAY OF TYPE 2 DIABETES V. DIABETES CARE

A. Initial evaluation B. Management C. Glycemic control

1. Assessment of glycemic control a. Glucose monitoring b. HbA1c

2. Glycemic goals in adults D. Pharmacological and overall approaches to treatment

1. Insulin therapy for type 1 diabetes 2. Pharmacological therapy for hyperglycemia in type 2 diabetes

E. MNT - medical nutrition therapy F. Diabetes self-management education and support G. Physical activity H. Psychosocial assessment and care I. When treatment goals are not met J. Intercurrent illness K. Hypoglycemia L. Bariatric surgery M. Immunization

VI. PREVENTION AND MANAGEMENT OF DIABETES COMPLICATIONS

A. Cardiovascular disease (CVD) 1. Hypertension/blood pressure control 2. Dyslipidemia/lipid management 3. Antiplatelet agents 4. Smoking cessation 5. CHD screening and treatment

B. Nephropathy screening and treatment

217

C. Retinopathy screening and treatment D. Neuropathy screening and treatment E. Foot care

VII. ASSESSMENT OF COMMON COMORBID CONDITIONS VIII. DIABETES CARE IN SPECIFIC POPULATIONS

A. Children and adolescents 1. Type 1 diabetes

a. Glycemic control b. Screening and management of chronic complications in

children and adolescents with type 1 diabetes i. Nephropathy ii. Hypertension iii. Dyslipidemia iv. Retinopathy v. Celiac disease vi. Hypothyroidism

c. Self-management d. School and day care e. Transition from pediatric to adult care

2. Type 2 diabetes 3. Monogenic diabetes syndromes

B. Preconception care C. Older adults D. Cystic fibrosis–related diabetes

IX. DIABETES CARE IN SPECIFIC SETTINGS A. Diabetes care in the hospital

1. Glycemic targets in hospitalized patients 2. Antihyperglycemic agents in hospitalized patients 3. Preventing hypoglycemia 4. Diabetes care providers in the hospital 5. Self-management in the hospital 6. MNT in the hospital 7. Bedside blood glucose monitoring 8. Discharge planning and DSME

B. Diabetes and employment C. Diabetes and driving D. Diabetes management in correctional institutions

X. STRATEGIES FOR IMPROVING DIABETES CARE Objective 1: Optimize provider and team behavior Objective 2: Support patient behavior change Objective 3: Change the system of care

(American Diabetes Association, 2015)

218

Appendix E

Interview Participants

Specialty Year Finished

Residency

Current Practice Location

Practice Setting

Date Duration (hh:mm:ss)

P01 Family Medicine

1998 Manchester, NH

FQHC 2.25.2014 46:30

P02 Internal Medicine

1995 Hartford, CT

Clinic of Public Hospital

2.25.2014 45:44

P03 Internal Medicine

2012 E. Lansing, MI

Clinic of Public Hospital

2.28.2014 34:50

P04 Internal Medicine

2015 Detroit, MI Clinic of Public Hospital

3.20.2014 57:37

P05 Endocrinology 2012 Detroit, MI Clinic of Public Hospital

3.31.2014 49:46

P06 Family Medicine

2007 Ann Arbor, MI

Clinic of Teaching Hospital

5.29.2014 1:08:39

P07 Internal Medicine

2004 Ann Arbor, MI

University Teaching Hospital

6.9.2014 1:01:10

P08 Internal Medicine

2000 E. Lansing, MI

University Teaching Hospital

6.23.2014 56:39

P09 Internal Medicine

1991 Dallas, TX Clinic of Private Hospital

6.25.2014 50:19

P10 Family Medicine

2007 Dallas, TX Community Clinic

6.25.2015 1:00:27

P11 Internal Medicine

2010 Dallas, TX Community Clinic

6.25.2014 54:45

P12 Family Medicine

1984 Boston, MA

Community Clinic

6.27.2014 1:12:23

P13 Internal Medicine

2009 Ann Arbor, MI

VA, University Teaching Hospital

7.10.2014 1:10:30

P14 Family Medicine

2010 Ann Arbor, MI

VA, University Teaching Hospital

7.15.2014 1:04:46

P15 Family Medicine

2011 Ann Arbor, MI

VA, University Teaching Hospital

7.15.2014 1:20:49

P16 Family Medicine

1998 Ann Arbor, MI

VA, University Teaching Hospital

7.29.2014 1:05:21

P17 Family Medicine

2003 Portage, IN FQHC 1.16.2015 49:02

219

Appendix F

Emails Used for Survey Distribution

Genesis Physicians Group

Genesis Physicians Group – Initial Email To PODs - 2.5.2015 To:Genesis PODS Send Date:February 5, 2015 @ 10:00 AM Link Type:Individual Link Show Link Options From Address:csenteio@umich.edu From Name:AMD POD leader Reply-To Email:csenteio@umich.edu Subject:GAPN POD Communication – Online Survey – Response Requested

Dr. $[m://FirstName] $[m://LastName], I want to introduce myself to you as the GAPN Associate Medical Director (AMD) for your practice. Along with you, approximately 15 other PCPs have been clustered into a “POD” according to geographic region. Currently GAPN has 6 PODs, totaling approximately 100 PCPs.

• GAPN is working diligently to engage each Primary Care Physician member in communicating KEY information thru their GAPN Associate Medical Director. To get started, I wanted to ask that you take 10 minutes and complete this online survey concerning type 2 diabetes care, even if you may have completed it in the past couple of months. Follow this link to the Survey: $[l://SurveyLink?d=Take the Survey] Or copy and paste the URL below into your internet browser: $[l://SurveyURL] The survey is an important first step and will help GAPN understand how its PCP members consider psychosocial information in making type 2 diabetes care decisions, and under which circumstances they use that information. Additionally, the survey will instruct us with how effective email communication works for important GAPN messages. We would like to learn:

• The level of physician engagement between the physicians and the Associate Medical Director (AMD) • GAPN’s effectiveness in responding to information physicians share about their practice patterns and needs

I ask that you please complete the survey in the next 5-7 days, as we intend to communicate the results during the first POD meeting of 2015. I will track the completion percentage of my POD and, as needed, may reach out to you with a gentle reminder to complete the survey. Thank you in advance for your involvement in GAPN, and I look forward to working with you in 2015. Sincerely, Fname, LName GAPN, Associate Medical Director Follow the link to opt out of future emails: $[l://OptOutLink?d=Click here to unsubscribe]

220

Genesis Physicians Group – Reminder Fax - 3.6.2015

March 6, 2015 Re: Reminder – Request to Complete Diabetes Care Survey Dear GAPN POD member, This is a reminder to please complete the survey emailed to you from your GAPN Associate Medical Director POD back on February 5th. Below is a copy of the email. Please check your SPAM Folder if you have not seen this email. We will send a reminder email on Friday, March 6th at 5pm. The survey is only accessible via the link in the survey email. Sincerely, Jim Walton President & CEO Dr. [FirstName] [LastName], I want to introduce myself to you as the GAPN Associate Medical Director (AMD) for your practice. Along with you, approximately 15 other PCPs have been clustered into a “POD” according to geographic region. Currently GAPN has 6 PODs, totaling approximately 100 PCPs.

• GAPN is working diligently to engage each Primary Care Physician member in communicating KEY information thru their GAPN Associate Medical Director.

To get started, I wanted to ask that you take 10 minutes and complete this online survey concerning type 2 diabetes care, even if you may have completed it in the past couple of months.

Follow this link to the Survey: $[l://SurveyLink?d=Take the Survey] Or copy and paste the URL below into your internet browser: $[l://SurveyURL]

The survey is an important first step and will help GAPN understand how its PCP members consider psychosocial information in making type 2 diabetes care decisions, and under which circumstances they use that information. Additionally, the survey will instruct us with how effective email communication works for important GAPN messages. We would like to learn:

• The level of physician engagement between the physicians and the Associate Medical Director (AMD)

• GAPN’s effectiveness in responding to information physicians share about their practice patterns and needs

I ask that you please complete the survey in the next 5-7 days, as we intend to communicate the results during the first POD meeting of 2015. I will track the completion percentage of my POD and, as needed, may reach out to you with a gentle reminder to complete the survey. Thank you in advance for your involvement in GAPN, and I look forward to working with you in 2015. Sincerely, GAPN, Associate Medical Director

221

Genesis Physicians Group – Reminder Email - 3.6.2015 To:Genesis PODS Send Date:March 6, 2015 @ 6:00 PM Link Type:Individual Link Show Link Options From Address:csenteio@umich.edu From Name:AMD POD leader Reply-To Email:csenteio@umich.edu Subject:GAPN POD Communication – Online Survey – Response Requested

Dear GAPN POD member, This is a reminder to please complete the survey emailed to you from your GAPN Associate Medical Director POD back on February 5th. Below is a copy of the email, with a link to access the survey. Sincerely, Jim Walton President & CEO Dr. $[m://FirstName] $[m://LastName], I want to introduce myself to you as the GAPN Associate Medical Director (AMD) for your practice. Along with you, approximately 15 other PCPs have been clustered into a “POD” according to geographic region. Currently GAPN has 6 PODs, totaling approximately 100 PCPs.

• GAPN is working diligently to engage each Primary Care Physician member in communicating KEY

information thru their GAPN Associate Medical Director. To get started, I wanted to ask that you take 10 minutes and complete this online survey concerning type 2 diabetes care, even if you may have completed it in the past couple of months. Follow this link to the Survey: $[l://SurveyLink?d=Take the Survey] Or copy and paste the URL below into your internet browser: $[l://SurveyURL] The survey is an important first step and will help GAPN understand how its PCP members consider psychosocial information in making type 2 diabetes care decisions, and under which circumstances they use that information. Additionally, the survey will instruct us with how effective email communication works for important GAPN messages. We would like to learn:

• The level of physician engagement between the physicians and the Associate Medical Director (AMD) • GAPN’s effectiveness in responding to information physicians share about their practice patterns and needs

I ask that you please complete the survey in the next 5-7 days, as we intend to communicate the results during the first POD meeting of 2015. I will track the completion percentage of my POD and, as needed, may reach out to you with a gentle reminder to complete the survey. Thank you in advance for your involvement in GAPN, and I look forward to working with you in 2015. Sincerely, GAPN, Associate Medical Director Follow the link to opt out of future emails: $[://OptOutLink?d=Click here to unsubscribe]

222

Genesis Physicians Group – Reminder Email from J. Walton - 3.11.2015 From: Jim Walton Sent: Wednesday, March 11, 2015 1:46 PM To: GAPNPODLEADERS Subject: AMD Alert All, I continue to work on discovering how best to communicate with the POD members. The graphics below reveal how your colleagues are responding to a request from GAPN and the AMDs. • Could you please take a look and see if there is anything you can do to call, email or fax your POD members that have NOT responded to the communication? • The last idea I have, is to try mailing a letter and asking the physicians to complete the survey to see what the rate of response might be. What are your thoughts? Please send me an email to see how to proceed. Thanks Jim Jim Walton President & CEO Genesis Physicians Group 5501 LBJ Freeway, Suite 950 Dallas, TX 75240 Direct: 972-419-0047 Cell: 214-399-8993 Fax: 972-239-3734 Learn more about Genesis: www.genesisdocs.org www.genesisaco.com CONFIDENTIALITY NOTICE TO RECIPIENT: This transmission contains confidential information belonging to the sender that is legally privileged and proprietary and may be subject to protection under the law, including the Health Insurance Portability and Accountability Act (HIPAA). If you are not the intended recipient of this e-mail, you are prohibited from sharing, copying, or otherwise using or disclosing its contents. If you have received this e-mail in error, please notify the sender immediately by reply e-mail and permanently delete this e-mail and any attachments without reading, forwarding or saving them. Thank you.

223

Genesis Physicians Group – Reminder Letter from J. Walton - 4.7.2015

[First Name] [Last Name] [Address 1] [Address 2] [City, State Zip] April 7, 2015 Re: Reminder – Request to Complete Diabetes Care Survey Emailed to You Dear [First Name] [Last Name], This is a reminder to please complete the survey emailed to you at [email address] from your GAPN Associate Medical Director on February 5th. A reminder email was sent on Friday, March 6th at 5pm from this address csenteio@umich.edu. The diabetes care survey is an important initial step to engage each Primary Care Physician `member in communicating KEY information thru your GAPN Associate Medical Director. The survey is only accessible via the link in the email, please check your spam folder if it is not in your inbox. If you need the email resent, please contact Charles Senteio at csenteio@umich.edu. Sincerely, Jim Walton President & CEO

224

North Texas Nurse Practitioners (NTNP)

NTNP – Initial Email – 11.20.2014 To:Providers (NTNP)_11.20.2014 Send Date:November 20, 2014 @ 10:04 PM Link Type:Individual Link Show Link Options From Address:csenteio@umich.edu From Name:Badia Harlin Reply-To Email:csenteio@umich.edu Subject:Psychosocial Factors in Diabetes Care [Provider Survey] Dear Colleagues with the North Texas Nurse Practitioners: I am writing to ask for your help in completing a 10-minute, online provider survey. I am supporting a University of Michigan health informatics study focusing on the psychosocial patient factors you may consider when caring for your type-2 diabetes patients in the outpatient setting. As part of this effort, this survey is designed to collect provider perspectives on psychosocial information. The overall goal is to enhance personalized care by increasing access to patient psychosocial information. This survey is critical to helping understand how providers may consider this information, and under which circumstances. The study aim is to better support care teams as they make, and influence, care decisions such as "the right drug at the right time," and making referrals and recommendations. As you know, several different provider roles can be involved in these important decisions. Your input is critically important. I would greatly appreciate you taking approximately 10 minutes of your time to complete the survey. If you have any level of experience seeing adult type-2 patients in the outpatient setting, and are willing to fill out the survey, please click on this link to take the survey [SurveyLink?d=Take the Survey] Or copy and paste the URL below into your internet browser: [SurveyURL] If you have any questions on the survey, or the study, please contact Charles Senteio, the study Principal Investigator, at csenteio@umich.edu. Thank you, Badia Harlin, DNP, FNP-c Follow the link to opt out of future emails: [l://OptOutLink?d=Click here to unsubscribe]

225

Michigan Association of Diabetes Educators (Diabetes Educators)

Diabetes Educators – Initial Email – 12.2.2014 Dear Colleagues at the Michigan Branch of the American Association of Diabetes Educators: I am writing to ask for your help in completing a 10-minute, online provider survey. I am supporting a University of Michigan health informatics study focusing on the psychosocial patient factors you may consider when caring for your type-2 diabetes patients in the outpatient setting. As part of this effort, we are using this survey to collect provider perspectives on psychosocial information. Our overall goal is to enhance personalized care by increasing access to patient psychosocial information. This survey is critical to helping us understand how providers may consider this information, and under which circumstances. Our aim is to better support care teams as they make, and influence, care decisions such as "the right drug at the right time," and making referrals and recommendations. As you know, several different provider roles can be involved in these important decisions. Therefore we are currently surveying physicians, physician assistants, nurse practitioners, registered nurses, and clinical pharmacists. Your input as Diabetes Educators is critically important. I would greatly appreciate you taking approximately 10 minutes of your time to complete the survey. If you have any level of experience seeing adult type-2 patients in the outpatient setting, and are willing to fill out the survey, please click on the link below or paste the address in your browser. https://umich.qualtrics.com/SE/?SID=SV_41GUqpPUIJtLiOF Please contact Charles Senteio, the study Principal Investigator, at csenteio@umich.edu if you have specific questions regarding the study, or the survey. Thank you, Martha M. Funnell, MS, RN, CDE Associate Research Scientist Department of Learning Health Sciences Michigan Diabetes Research and Training Center 1111 E. Catherine St. 225 Victor Vaughan Ann Arbor, MI 48109-2054 T-(734) 936-9237 F-(734) 936-1641 mfunnell@umich.edu

226

Diabetes Educators – Reminder Email – 12.182014 Dear Colleagues at the Michigan Branch of the American Association of Diabetes Educators (or MODE for those of you in my age bracket): This is a reminder email regarding an online survey I sent back on December 2nd. The survey is part of a UM health informatics study concerning psychosocial factors and diabetes care. Thanks to those of you who have submitted your responses! Our preliminary findings based on what you provided are very helpful to our goal of understanding how patient psychosocial factors impact diabetes care. We are currently 16% toward our response goal.

For those of you with any level of experience seeing adult type-2 patients in the outpatient setting who have not yet taken the opportunity to complete the survey, please consider doing so by accessing this link: https://umich.qualtrics.com/SE/?SID=SV_41GUqpPUIJtLiOF I am also including below additional detail on our study objectives, which was also sent in my original email. Martha M. Funnell, MS, RN, CDE Associate Research Scientist Department of Learning Health Sciences Michigan Diabetes Research and Training Center 1111 E. Catherine St. 225 Victor Vaughan Ann Arbor, MI 48109-2054 T-(734) 936-9237 F-(734) 936-1641 mfunnell@umich.edu

227

Diabetes Educators – Reminder Email – 1.5.2015 Dear Colleagues at the Michigan Branch of the American Association of Diabetes Educators (or MODE for those of you in my age bracket): This is another reminder email regarding an online survey I sent back on December 2nd. The survey is part of a UM health informatics study concerning psychosocial factors and diabetes care. Thanks to those of you who have submitted your responses! Our preliminary findings based on what you provided are very helpful to our goal of understanding how patient psychosocial factors impact diabetes care. We are currently 42% toward our response goal. For those of you with any level of experience seeing adult type-2 patients in the outpatient setting who have not yet taken the opportunity to complete the survey, please consider doing so by accessing this link: https://umich.qualtrics.com/SE/?SID=SV_41GUqpPUIJtLiOF I am also including below additional detail on our study objectives, which was also sent in my original email.

I am supporting a University of Michigan health informatics study focusing on the psychosocial patient factors you may consider when caring for your type-2 diabetes patients in the outpatient setting. As part of this effort, we are using this survey to collect provider perspectives on psychosocial information. Our overall goal is to enhance personalized care by increasing access to patient psychosocial information.

228

This survey is critical to helping us understand how providers may consider this information, and under which circumstances. Our aim is to better support care teams as they make, and influence, care decisions such as "the right drug at the right time," and making referrals and recommendations. As you know, several different provider roles can be involved in these important decisions. Your input as Diabetes Educators is critically important. I would greatly appreciate you taking approximately 10 minutes of your time to complete the survey. Please contact Charles Senteio, the study Principal Investigator, at csenteio@umich.edu if you have specific questions regarding the study, or the survey. Martha M. Funnell, MS, RN, CDE Associate Research Scientist Department of Learning Health Sciences Michigan Diabetes Research and Training Center 1111 E. Catherine St. 225 Victor Vaughan Ann Arbor, MI 48109-2054 T-(734) 936-9237 F-(734) 936-1641 mfunnell@umich.edu

229

Diabetes Educators – Reminder Email – 1.23.2015 Dear Colleagues at the Michigan Branch of the American Association of Diabetes Educators (or MODE for those of you in my age bracket): This is another reminder email regarding an online survey I sent at the end of December. The survey is part of a UM health informatics study concerning psychosocial factors and diabetes care. Thanks to those of you who have submitted your responses! Our preliminary findings based on what you provided are very helpful to our goal of understanding how patient psychosocial factors impact diabetes care. We are currently 56% toward our response goal! For those of you with any level of experience seeing adult type-2 patients in the outpatient setting who have not yet taken the opportunity to complete the survey, please consider doing so by accessing this link: https://umich.qualtrics.com/SE/?SID=SV_41GUqpPUIJtLiOF I am also including below additional detail on our study objectives, which was also sent in my original email.

I am supporting a University of Michigan health informatics study focusing on the psychosocial patient factors you may consider when caring for your type-2 diabetes patients in the outpatient setting. As part of this effort, we are using this survey to collect provider perspectives on psychosocial information. Our overall goal is to enhance personalized care by increasing access to patient psychosocial information. This survey is critical to helping us understand how providers may consider this information, and under which circumstances. Our aim is to better support care teams as they make, and influence, care decisions such as "the right drug at the right time," and making referrals and recommendations. As you know, several different provider roles can be involved in these important decisions. Your input as Diabetes Educators is critically important.

230

Martha M. Funnell, MS, RN, CDE Associate Research Scientist Department of Learning Health Sciences Michigan Diabetes Research and Training Center 1111 E. Catherine St. 225 Victor Vaughan Ann Arbor, MI 48109-2054 T-(734) 936-9237 F-(734) 936-1641 mfunnell@umich.edu

231

Appendix G

Semi-Structured Interview Guide

1. Describe your experience in treating adult, type 2 diabetes patients. [Specific Probes: Length of time treating adult T2DM patients? Current practice setting? Types of T2DM patient population(s) you have treated (e.g. rural/urban; income level; payor status; age, gender, comorbidity-which?]

2. Describe situations in which you strongly consider clinical practice guidelines? In which situations do you consider them less? [Specific Probes: When you do consider them less, what factors do you consider in establishing alternative treatment goals for specific patients, which we’ll refer to as ‘patient-centered care goals? What are some examples of these patient-centered care goals that may differ from CPGs, or those that may be organizational or practice goals (e.g. HbA1c of 7.8, instead of 7.0)? What patient characteristics trigger consideration?]

3. There is no one, universal definition of psychosocial factors (Martikainen et. al., 2002). For the purposes of this study, we define psychosocial factors as those that are not genetic, nor directly driven by comorbid conditions; they are psychological and social factors that affect patients, their families and health care providers (IOM, 2008). Describe the psychosocial factors you think are important in making care decisions (e.g. decisions in pursuit of treatment goals which differ from CPGs, referral to support services). How do you consider these factors? What additional psychosocial information would you like to use? Why don’t you use it? [Guide for Probes: In RWJ’s 2011 Health Care’s Blind Side Survey of 1,000 physicians they identified social factors that affect health outcomes: inadequate housing, un-/under-employment, access barriers to transportation, food, or other neighborhood deficiencies. Also, family factors like family conflict and stress, high levels of cohesion and organization, good communication are associated with T2DM regimen adherence and metabolic control (Delamater et. al., 2001).]

4. Describe the situations in which you think it is more important to consider psychosocial factors. For example, for certain clinical situations (e.g. A1c spike, certain patient populations, new patient)? Please describe specific psychosocial information considered more heavily, for specific situations. What role, if any, might psychosocial factors play for new patients, the first patient visit? [Guide for Probes: Example (if needed) patient cases – Mr. A has a history of family stability and employment while Mrs. B is faced with declining cognition and stressed caregiver - can help describe circumstance when certain patient situations trigger certain treatment alternatives (Hackel, 2013). Age – comorbid depression higher in diabetic women than men (Anderson, 2001), other factors could be: employment, housing stability, family support.]

232

5. How do you use this information in setting patient goals? [Guide for Probes: Where/how do you get information on these factors (EMR)? What are some barriers to accessing this information and using it? Which are currently accessible in the medical record? To what degree do you use this info that may be in the record? What additional information is needed?]

6. Where do you get psychosocial information (specific patient, and/or patient population/groups)? What informs your insight on psychosocial factors? Is there additional information you do not access? If so, how? If not, why not? If you were training new physicians, what psychosocial factors would you tell them they should consider? In which circumstances? [Guide for Probes: How do you know if a patient is uninsured, or experiencing low social support? How do you know if they are experiencing financial barriers to adherence? How do you know if they are experiencing stress due to neighborhood violence? How do you know if a patient may be experiencing barriers to access healthy foods? What information is currently in the medical record? What needs to be added? What are your perceptions of the accuracy of psychosocial (patient) data that is self-reported?]

7. How can we improve collection and use of psychosocial information in order to better support your care decisions? Would an index help you? How would you access and keep it current?

8. Demographics: Type of practice; Family or GP, Internal Medicine; When/where did you complete medical school? When did you complete residency? Type of residency? Where? Did you treat T2DM pts during residency? Where/how did you gather psychosocial insights gathered?

9. Is there anything else you’d like to tell me about your experience with psychosocial factors in treating adult, T2DM pts?

233

Appendix H

Survey Instrument Instructions This 10 minute survey is part of a University of Michigan health informatics study that seeks to understand and document how patient psychosocial factors influence outpatient care decisions for adult, type 2 diabetes mellitus (T2DM) patients. Your participation in the survey will help:

• understand the psychosocial factors that inform T2DM decisions at the point of care

• enable providers and developers of health informatics capabilities (i.e. clinical decision support

• systems, EHRs) to more effectively capture and use psychosocial information Demographics Indicate your role in treating adult, type 2 diabetes (T2DM) patients in the outpatient setting.

o Physician (MD or DO) o Nurse Practitioner (NP) o Physician Assistant (PA) o Registered Nurse (RN) o Clinical Pharmacist o Other

o I do not have experience treating adult, T2DM patients in the outpatient setting

Are you a Certified Diabetes Educator (CDE)?

o Yes o No

What is your clinical specialty? If you do not have a specific one, indicate the one in which you have the most experience with adult, T2DM patients.

234

(Options in Drop-down Menu) Have you had specific, formal training (e.g. courses in professional, graduate, post-graduate school, CEUs, etc.) in treating adult, T2DM patients?

o Yes o No

What is your age?

o Under 25 years old o 25-34 years old o 35-44 years old o 45-54 years old o 55-64 years old o 65-74 years old o 75-84 years old o 85 years old or older

Indicate your gender.

o Male o Female

Emergency Medicine Endocrinology Hematology/oncology Hospitalist Internal medicine Nephrology Obstetrics/gynecology Ophthalmology Pathology Pediatrics Podiatry Surgery Urology Other

235

Indicate your race/ethnic identity (select all that apply).

o White o Black or African American o Hispanic or Latino o Asian o Native American or Alaska Native o Native Hawaiian or Other Pacific Islander o Other

How much experience, not including formal clinical training, do you have treating adult, T2DM patients? Years (Whole Number) Months (Whole Number 1-11) Indicate the practice setting, including formal clinical training, where you have treated adult, T2DM patients (select all that apply).

o Community Health Center o Federally Qualified Health Center (FQHC) o Free clinic o Home Health / Home-based care o Outpatient clinic of a hospital system o Outpatient clinic of a teaching hospital o Clinic affiliated with the Veterans Administration (VA) system o Indian Health Services (IHS) o A small private group practice (5 or fewer physicians) o A large private group practice (6 or more physicians) o Other

Have you had a formal instructor role in training other providers in the care of adult, T2DM patients (Example roles include: attending physician, instructor for CEUs, instructor for Diabetes Intensive Training Course, etc.)?

o Yes o No

236

Clinical Decisions There are 4 groups of psychosocial factors that can influence self-care behavior: I. Sociodemographic (e.g. financial demands, culture) II. Psychological (e.g. mental health status) III. Social relationship (e.g. social support) IV. Neighborhood (e.g. housing and food security) Although psychosocial factors are currently acknowledged to affect treatment behavior, little is known about how providers consider psychosocial information as they make diabetes care decisions. The next questions concern the circumstances when psychosocial factors may be considered, and the specific T2DM decisions that they may influence. In which circumstances are psychosocial factors, as a group, important to consider when making clinical decisions for adult, T2DM patients in the outpatient setting (select all that apply)? If you believe psychosocial factors are important in all circumstances, select the "In all circumstances" option near the bottom of the list.

o Seeing a new patient o Seeing a work-in patient (i.e. seeing a colleague's patient) o Change in patient health status (e.g. spike in HbA1c, additional diagnosis) o Patient with persistent, low treatment adherence o Seeing a patient from low-resourced areas o Seeing a patient with multiple chronic conditions o Seeing a patient with a diagnosed mental health condition o Seeing a patient with undiagnosed mental health issues o In all circumstances o Other

237

In your treatment of adult, T2DM patients in the outpatient setting, how often do psychosocial factors influence your decisions -- or influence your input on clinical decisions in the following areas (place mouse over the decisions for examples)?

Always

Often

Sometimes

Rarely

Never

N/A – I haven’t

made nor had input on decisions in

this area Target level of control O O O O O O

Medications O O O O O O Making Referrals O O O O O O

Other Decisions O O O O O O

For the following decisions associated with determining level of control, how often do psychosocial factors, as a group, influence these decisions -- or influence your input on them?

Always

Often

Sometimes

Rarely

Never

N/A – I haven’t

made nor had input on decisions in

this area Establish target goal for

blood glucose

O O O O O O

Establish target goal for HbA1c

O O O O O O

Incorporate input from patient in setting goal

O O O O O O

Other Target Goal Decision(s) O O O O O O

238

For the following medication decisions, how often do psychosocial factors, as a group, influence these decisions -- or influence your input on them?

Always

Often

Sometimes

Rarely

Never

N/A – I haven’t

made nor had input on decisions in

this area Select a specific

medication (e.g. long acting insulin or short

acting insulin)

O O O O O O

Select a brand, or a generic medication

O O O O O O

Start a patient on first oral diabetes medication

O O O O O O

Adjust oral diabetes medication dosage

O O O O O O

Add an additional (2nd, 3rd, 4th, etc.) oral diabetes

medication

O O O O O O

Start a patient on injectable insulin

O O O O O O

Adjust injectable insulin dosage

O O O O O O

Start a patient on non-insulin injectable

diabetes medication (e.g. GLP-1)

O O O O O O

Adjust non-insulin injectable diabetes

medication (e.g. GLP-1)

O O O O O O

Reduce complexity of medication regimen

O O O O O O

Other Medication Decision(s)

O O O O O O

239

For the following referral decisions, how often do psychosocial factors, as a group, influence these decisions -- or influence your input on them?

Always

Often

Sometimes

Rarely

Never

N/A – I haven’t

made nor had input on decisions in

this area Refer a patient to specialty

care (e.g., endocrine, mental health, etc.)

O O O O O O

Refer a patient to support services outside the

organization (e.g., social work, housing,

transportation, etc.)

O O O O O O

Refer a patient to support services within the

organization (e.g., social work, housing,

transportation, etc.)

O O O O O O

Refer a patient to diabetes education

O O O O O O

Refer a patient to a dietitian and/or nutritional

information\

O O O O O O

Other Referral Decision(s) O O O O O O

For the following recommendation decisions, how often do psychosocial factors, as a group, influence these decisions -- or influence your input on them?

Always

Often

Sometimes

Rarely

Never

N/A – I haven’t

made nor had input on decisions in

this area Make dietary

recommendations O O O O O O

240

Make physical activity recommendations

O O O O O O

Recommend a patient's caregivers understand

what is required of patient

O O O O O O

Frequency of clinical visits O O O O O O

Other Recommendation Decision(s)

O O O O O O

In making or informing these same decisions, how often do you feel that you have the psychosocial information that you need (place mouse over the decisions for examples)?

Always

Often

Sometimes

Rarely

Never

N/A – I haven’t

made nor had input on decisions in

this area Target level of control

O O O O O O

Medications

O O O O O O Making Referrals

O O O O O O

Making Recommendations

O O O O O O Other Decision(s) O O O O O O

Individual Psychosocial Factors & Sources There are 4 groups of psychosocial factors that can influence self-care behavior: I. Sociodemographic (e.g. financial demands, culture) II. Psychological (e.g. mental health status) III. Social relationship (e.g. social support) IV. Neighborhood (e.g. housing and food security)

241

From the following sociodemographic psychosocial factors, indicate their importance in treating adult, T2DM patients (place mouse over each factor for a description).

Very Important

Important

Neither Important

nor Unimportant

Unimportant

Very Unimportant

Financial Strain

O O O O O

Employment

O O O O O

Payor status/Type of

insurance

O O O O O

Culture and spirituality

O O O O O

Other responsibilities

O O O O O

Level of education

O O O O O

Literacy

O O O O O

Country of origin

O O O O O

Level of English proficiency

O O O O O

Sources of Sociodemographic Psychosocial Information For the sociodemographic psychosocial information you indicated is Extremely Important or Very Important, select the source(s) you use to gather this information about the patient (select all sources that apply).

Patient

Family/ Caregivers

Other providers

or members of the care

team

Electronic

Health Record (EHR)

Other – Please specify

No Reliable Source

Financial Strain

O O O O O

Employment

O O O O O

Payor status/Type of

insurance

O O O O O

242

Culture and spirituality

O O O O O

Other responsibilities

O O O O O

Level of education

O O O O O

Literacy

O O O O O

Country of origin

O O O O O

Level of English proficiency

O O O O O

There are 4 groups of psychosocial factors that can influence self-care behavior: I. Sociodemographic (e.g. financial demands, culture) II. Psychological (e.g. mental health status) III. Social relationship (e.g. social support) IV. Neighborhood (e.g. housing and food security) From the following psychological psychosocial factors, indicate their importance in treating adult, T2DM patients (place mouse over each factor for a description).

Very Important

Important

Neither Important

nor Unimportant

Unimportant

Very Unimportant

Mental health status

O O O O O

Life Stressors

O O O O O

TDM perceptions-

beliefs

O O O O O

Health Literacy

O O O O O

Health Numeracy

O O O O O

Sources of Psychological Psychosocial Information

243

For the psychological psychosocial information you indicated is Extremely Important or Very Important, select the source(s) you use to gather this information about the patient (select all sources that apply).

Patient

Family/ Caregivers

Other providers

or members of the care

team

Electronic

Health Record (EHR)

Other – Please specify

No Reliable Source

Mental health status

O O O O O O

Life Stressors

O O O O O O

TDM perceptions-

beliefs

O O O O O O

Health Literacy

O O O O O O Health

Numeracy

O O O O O O

There are 4 groups of psychosocial factors that can influence self-care behavior: I. Sociodemographic (e.g. financial demands, culture) II. Psychological (e.g. mental health status) III. Social relationship (e.g. social support) IV. Neighborhood (e.g. housing and food security) From the following social relationships/living conditions psychosocial factors, indicate their importance in treating adult, T2DM patients (place mouse over each factor for a description).

Very Important

Important

Neither Important

nor Unimportant

Unimportant

Very Unimportant

Social Support

O O O O O

Threat of Violence – from

abusive relationships

O O O O O

Threat of Violence – from

the community

O O O O O

244

Sources of social relationships/living conditions Psychosocial Information For the psychological psychosocial information you indicated is Extremely Important or Very Important, select the source(s) you use to gather this information about the patient (select all sources that apply).

Patient

Family/ Caregivers

Other providers

or members of the care

team

Electronic

Health Record (EHR)

Other – Please specify

No Reliable Source

Social Support

O O O O O O

Threat of Violence – from

abusive relationships

O O O O O O

Threat of Violence – from

the community

O O O O O O

There are 4 groups of psychosocial factors that can influence self-care behavior: I. Sociodemographic (e.g. financial demands, culture) II. Psychological (e.g. mental health status) III. Social relationship (e.g. social support) IV. Neighborhood (e.g. housing and food security) From the following neighborhood/community psychosocial factors, indicate their importance in treating adult, T2DM patients (place mouse over each factor for a description).

Very Important

Important

Neither Important

nor Unimportant

Unimportant

Very Unimportant

Patient's Rural/ Urban/ Suburban residence setting

O O O O O

Neighborhood residence

O O O O O

Housing security O O O O O

245

Food security

O O O O O

Access to transportation

O O O O O

Access to places to exercise

O O O O O

Literacy

O O O O O

Country of origin

O O O O O

Level of English proficiency

O O O O O

Sources of neighborhood/community Psychosocial Information For the neighborhood/community psychosocial information you indicated is Extremely Important or Very Important, select the source(s) you use to gather this information about the patient (select all sources that apply).

Patient

Family/ Caregivers

Other providers

or members of the care

team

Electronic

Health Record (EHR)

Other – Please specify

No Reliable Source

Patient's Rural/ Urban/

Suburban residence setting

O O O O O

Neighborhood residence

O O O O O

Housing security

O O O O O

Food security

O O O O O Access to

transportation

O O O O O

Access to places to exercise

O O O O O

Accuracy I am confident in the accuracy of psychosocial information I get from the following sources.

246

Always

Often

Sometimes

Rarely

Never

N/A – I haven’t

made nor had input on decisions in

this area Patient – via consultation

or interview

O O O O O O

Patient – self-reported via Screening Tools/Forms

O O O O O O

Family/ caregivers O O O O O O

Other providers or members of the care team

O O O O O O

Electronic Health Record (EHR)

O O O O O O

Other – please specify O O O O O O

EHR Questions These questions concern the Electronic Health Record (EHR). I believe the following EHR tools support the documentation of pertinent psychosocial information.

Always

Often

Sometimes

Rarely

Never Templates

O O O O O Data Fields

O O O O O Free Text

O O O O O I am confident in my ability to use the following EHR tools to document pertinent psychosocial information.

247

Always

Often

Sometimes

Rarely

Never Templates

O O O O O Data Fields

O O O O O Free Text

O O O O O I believe the following EHR tools support the retrieval of pertinent psychosocial information.

Always

Often

Sometimes

Rarely

Never Templates

O O O O O Data Fields

O O O O O Free Text

O O O O O I am confident in my ability to use the following EHR tools to retrieve pertinent psychosocial information.

Always

Often

Sometimes

Rarely

Never Templates

O O O O O Data Fields

O O O O O Free Text

O O O O O The EHR(s) my organization(s) use is/are (select all that apply).

O ALHTA (Department of Defense)

248

O AllScripts

O athenaHealth

O Benchmark

O Care360

O CareCloud

O Centricity (GE)

O Cerner (PowerChart)

O E-MDs

O eClinicalWorks

O Epic (MyChart or MiChart)

O Kareo

O MedicsDocAssistant

O Meditech

O Meditouch

O Mercury Medical (CrisSoft)

O Microsoft Dynamics GP

O NueMD

O Office Practicum

O PrognoCIS

O VelociDoc (Practice Velocity)

O VistA (Veterans Administration)

O WRS Health

O An "in house" system specific to my organization which is not commercially available

O Other

O Not sure

249

These last 2, optional entries give you the opportunity, if you wish, to describe situations when psychosocial information influenced a specific decision, or your input on a decision. Describe an interaction with an adult, T2DM patient in which you learned later about a psychosocial factor that would have changed earlier decisions, or your input in earlier decisions, regarding that patient's treatment plan. Describe an interaction with an adult, T2DM patient in which you modified that patient's treatment plan, or provided input to modify the plan, because of psychosocial factors. Use the forward button to submit all answers and complete the survey. Thank you!

250

Appendix I Code Book for Physician Interviews

1. Add’l PI Required – “Additional Psychosocial Information Required” – expression of desire for additional psychosocial information, as it relates to making or influencing a clinical care decisions.

2. Adherence barriers – general barriers to self-care that are not explicitly expressed as psychosocial barriers a. Psychosocial – barriers to self-care explicitly expressed as driven by

psychosocial factors (e.g. financial barriers that prevent access to medications, neighborhood violence issues that prevent individuals from exercising).

3. Assessing adherence – describing perceived drivers of behavior that may differ from recommended self-care practices, or is inconsistent with healthy self-care behavior

4. DM Self-Care –various activities consistent with the recommended diabetes care regimen, including dietary habits, medication behavior, physical activity, and attending clinical appointments

5. EHR – references to the electronic health record (EHR), clinical record, or the record. Includes description of content of what the record contains, and how it is used a. Capture PI – using the EHR to store psychosocial information (PI). b. Use of PI – using EHR to document or use psychosocial information (PI) c. Accuracy of info capture – perspectives on the accuracy of information

captured in the EHR. d. Differences in inpatient v outpatient – references to differences in

capturing, or using, health information in the inpatient and outpatient setting. Includes references to information entered by other providers associated with inpatient or outpatient care (e.g. outpatient notes)

e. EHR Barriers to Capture PI – describing circumstances or situations when barriers to capture psychosocial information in the EHR are encountered

f. Improving EHR in PI capture and use – describing improvements to the EHR to better facilitate the capture and use of psychosocial information

6. How decisions are made – general references to how clinical care decisions are made, considering psychosocial factors. Includes influencing clinical decisions. a. Decisions – references to specific clinical decisions, include medications,

referrals (e.g. social work/social support, specialty care), recommendations (diet, physical activity, etc.)

b. DM guidelines – references to clinical decisions considering diabetes guidelines, for HbA1c or fasting blood sugar (FBS)

c. New Patients – describing making clinical decisions for new patients, specifically patients that the interview participant has not treated but may not be new to the clinical site and/or healthcare system

d. Overtreatment concerns – decisions made considering “too much” care, overtreatment

251

e. Risk-Life Expectancy – reference to making clinical decisions considering risk, including life expectancy of the patient

f. Pt. Engagement – describing making clinical decisions purposively and intentionally including direct input from the patient.

g. Newly diagnosed – making clinical decisions for patients newly diagnosed with diabetes

7. Limited Clinical Resources – references to low availability of clinical resources, including time, for the interview participant or for other members of the care team (e.g. nurses who may not have time for additional activities, increasing patient caseload that strains resources)

8. PFs – general references to psychosocial factors a. Social support – reference to social isolation, social connections, support

includes four dimensions: appraisal support, informational support, instrumental support and emotional support (Funnell, 2010)

b. Social Support NEGATIVE – references to aspects of patient’s social network that may act as barriers to healthy self-care activities, or overall wellness

c. Health Literacy – reference to measure of patients’ ability to read, comprehend, and act on medical instructions (Al Sayah, Williams, & Johnson, 2013; Schillinger, Grumbach, Piette, Wang, Osmond, Daher, Diaz Sullivan, et al., 2002)

d. Literacy – reference to the ability to use printed and written information to function in society, to achieve one’s goals, and to develop one’s knowledge and potential (Institute of Education Sciences, 2014)

e. Employment – reference to job demands that may influence self-care (e.g. work hours, type of job, ability to take time off for self-care)

f. Financial Stressors – references to patient’s lack of resources that impact access to food, safe housing, transportation, or medications

g. Payor status – reference to a patient’s insurance coverage, type of coverage, private insurance, government health insurance (Medicare, Medicaid, Military, State-specific plans, Indian Health Service) (U.S. Census Bureau, 2014)

h. Immigrant status – reference to a patient, or group of patients, who may not be documented U.S. citizens

i. Culture – reference to cultural norms and traditions which include dietary practices, faith beliefs, and practices (Kittler & Sucher, 1995b)

j. Transportation – reference to barriers which can lead to missed appointments, missed or delayed medication use (Syed et al., 2013)

k. Violence – reference to perceived patient perceiving threat of violence caused by violence-inducing or violence-protecting conditions (Sampson et al., 2005)

l. English proficiency – reference to patient’s ability to understand and speak English

m. Neighborhood – reference to patient’s residency setting, including physical aspects of a neighborhood that may influence a patient’s ability to purchase products (food), enable mobility, and interact and informally monitor another’s behavior (D. A. Cohen et al., 2003)

252

n. Access to Healthy Food – reference to access to fresh, healthy, and affordable food (Walker et al., 2010)

o. Importance of Psychosocial Info – description of general importance of psychosocial information

p. Mental health – reference to patient’s appearance; manner and approach; orientation, alertness, and thought processes; mood and affect (Sharma et al., 2004)

q. Stress – reference to negative patient’s events, chronic strains, traumas (Thoits, 2010)

r. DM perceptions-beliefs – description of patient perceptions of the relative quality-of-life effects of complications and treatments (Huang et al., 2007)

s. Education – reference to patient’s level of formal schooling t. Activities of daily living – reference to patient’s living situation and/or

demands that may impede self-care activities (e.g. cooking, shopping, caring for children)

u. Rural-Urban – reference to patient’s residency setting, that may influence health care utilization or access to care (Spoont et al., 2011)

v. Health numeracy – reference to degree to which individuals can obtain, process, and understand the basic quantitative health information and services they need to make appropriate health decisions (Al Sayah et al., 2013)

w. Faith – reference to patient’s faith that may influence self-care activities, and/or outlook

x. Housing security – reference to stable housing, access to affordable, safe housing (Johns Hopkins Center to Eliminate Cardiovascular Health Disparities, 2014b)

y. Food security – reference to access to fresh, healthy, and affordable food (Walker et al., 2010)

z. Violence-Domestic – reference to a pattern of coercion, physical abuse, sexual abuse, or threat of violence in personal relationships (Krug, 2002)

9. Psychosocial Index – reference to idea of development and use of a psychosocial index, to measure an individual’s psychosocial “health”, much like a Beck scale used for mental health/depression a. Pros – affirmative comments on the development and/or use of a psychosocial

index to inform clinical decisions b. Cons – negative comments on the development and/or use of a psychosocial

index to inform clinical decisions c. When used – comments on when a psychosocial index would be used d. How used – comments on how a psychosocial index would be used to inform

clinical decisions e. Items for Index – comments on what a psychosocial index should include

and/or incorporate 10. Pt. goals – general comments concerning patient goals as they pertain to the

diabetes care regimen a. Goal setting – references to setting specific goals b. Negotiating pt goals – references to negotiating goals with patients

253

11. Pt population – general comments about patient population currently served, or recently served a. Complexity – comments concerning complexity of a patient population b. Comorbidity – comments concerning comorbid conditions of a patient

population 12. Pt preferences – comments concerning patient preferences 13. Pt Story – references to the importance of understanding the patient’s situation

and/or circumstances 14. Source of insights on importance of PFs –general comments on the source of

insights on the importance of psychosocial factors influencing self-care and/or outcomes a. Clinical experience – reference to clinical experience providing insights on

the importance of psychosocial factors influencing self-care and/or outcomes b. Literature – reference to published literature providing insights on the

importance of psychosocial factors influencing self-care and/or outcomes c. Medical Training –reference to the physician’s medical training, including

residency, on the importance of psychosocial factors influencing self-care and/or outcomes

15. Source of PI – reference to the general sources of psychosocial information at it pertains to clinical decisions a. Pt-Doc Relationship – reference to the patient-doctor relationship being a key

driver in the source of psychosocial information b. Accessing PI – assessing the accuracy of psychosocial information c. Barriers to access PI – reference to barriers to access psychosocial

information d. Other providers – reference to other members of the care team as sources of

psychosocial information e. Caregivers – reference to caregivers, including family members, as sources of

psychosocial information f. EHR – reference to the EHR as a source of psychosocial information g. Home visit – reference to home visits as a source of psychosocial information h. Ideas to improve capture of PI – description of ideas to improve the capture

of psychosocial information 16. Training residents or med students – comment on training residents or medical

students on the importance of psychosocial information in making clinical decisions

17. Type of clinical practice setting – reference to type of clinical practice setting (e.g. community clinic, federally qualified health center)

18. When PFs Considered – reference to what circumstances and/or situations trigger the consideration of psychosocial factors

254

Appendix J

Principles of Survey Instrument Design

I asked two types of questions: 1) nonsensitive questions about behavior and possibly 2) attitude questions (Bradburn, Sudman, & Wansink, 2004). First, questions about behavior concerned what psychosocial factors (PFs) providers consider and during which situations (SFs) these factors are considered. As stated in Groves et al. (2011), I followed these guidelines concerning these types of (nonsensitive) behavior questions (p. 243):

1. With closed questions, include all reasonable possibilities as explicit response options

2. Use words that virtually all respondents will understand 3. Lengthen the questions by adding memory cues to improve recall 4. When forgetting is likely, use aided recall

Since I included questions about attitudes—common in survey methodology—I followed Groves et al. (2011) selected guidelines concerning designing questions soliciting attitudes (p. 248):

1. Avoid double-barreled questions 2. Measure the strength of the attitude, using separate items 3. Use bipolar items, except when they might miss important information 4. Carefully consider alternatives provided, as they influence the answers 5. Ask the general question first when asking general and specific questions

about a topic The first four guidelines concern question wording. Following these guidelines makes interpretation consistent across respondents. Double-barreled questions ask about two attitudes simultaneously. I sought attitudes about the two characteristics which are generally solicited via surveys, direction (affirmative or negative) and intensity (weak or strong). Assessing the direction is fairly straightforward; however, in soliciting strength I followed a summated rating scale (“strongly disagree” to “strongly agree”). I used a Likert scale, and referred to the literature on the granularity needed—and the interpretation of the (ordinal) data—prior to finalizing the survey instrument (Ogden & Lo, 2012; Spector, 2004; Tastle & Wierman, 2006). The last two guidelines reduce influence due to question ordering. General questions should precede more specific questions on similar topics. When asking questions about a specific topic (e.g., the influence of spousal support) respondents can show bias in subsequent, more general questions (e.g. the influence of family support). In this example, the question about family support should appear before the question about spousal support.

255

Appendix K Average Likert Scores of Psychosocial Factors

Complete list of the 23 psychosocial factors I investigated in the survey

instrument, listed by highest to lowest average score in the Likert scale.

Average Likert Scores of Individual Psychosocial Factors

Total (n = 164)

Primary Care Physicians

(n = 36)

NP & Diabetes Educators (n = 128) p value

Financial Strain 4.84 (383) 4.69 (.525) 4.88 (.323) .009 Mental Health Status 4.62 (.556) 4.28 (.701) 4.73 (.465) .001 Life Stressors 4.57 (.576) 4.25 (.649) 4.66 (.522) .000 Food Security 4.55 (.633)a 3.88 (.729)b 4.74 (.460)c .000 Social Support 4.53 (.572)a 4.20 (.678)d 4.62 (.504)e .000 Health Literacy 4.53 (.580) 4.14 (.683) 4.64 (.498) .000 T2DM Perceptions/Beliefs 4.52 (.660) 4.06 (.715) 4.65 (.583) .000

Access to Transportation 4.50 (.664)a 3.94 (.851)b 4.65 (.512)c .000 Threat of Violence - From Abusive Relationship(s)

4.48 (.810)a 4.03 (1.014)d 4.60 (.696)e .000

Literacy 4.47 (.650) 4.11 (.785) 4.57 (.571) .000 Health Numeracy 4.44 (.695)f 4.14 (.810)b 4.52 (.639) .004 Employment Status 4.39 (.678) 4.25 (.770) 4.43 (.648) .161 Payor Status 4.36 (.775) 4.22 (.866) 4.40 (.746) .229 Threat of Violence - From Community 4.35 (.880)a 3.83 (1.124)d 4.50 (.738)e .002

Housing Security 4.26 (.791)a 3.62 (.922)b 4.44 (.653)c .000 Other responsibilities 4.24 (.766) 3.75 (.731) 4.38 (.721) .000 Access to places to exercise 4.19 (.748)a 3.85 (.892)b 4.28 (.679)c .003

Culture and Spirituality 4.17 (.782)f 3.60 (.881)d 4.33 (.677) .000 Level of English Proficiency 4.15 (.764)f 3.92 (.874) 4.21 (.720)g .040

Level of Education 4.12 (.725) 3.94 (.674) 4.17 (.733) .084 Patient’s Rural/Urban/ Suburban residence setting

4.11 (.846)a 3.73 (.963)b 4.21 (.786)c .004

Neighborhood Residence 4.01 (.787)a 3.62 (.954)b 4.12 (.703)c .006 Country of Origin 3.55 (1.00)f 3.25 (.996) 3.64 (.989)g .040

Note. Respondents were asked to indicate the importance of psychosocial factors in making, or providing input into, diabetes care clinical decisions. Responses were captured in a Likert scale: 5 – Very Important, 4 – Important, 3 – Neither Important nor Unimportant. 2 – Unimportant, 1 – Very Unimportant. Standard deviations listed in parentheses. a n= 159. b n = 34. c n = 125. d n = 35. e n = 124. f n = 163. g n = 127.

256

Appendix L

Average Likert Scores of Psychosocial Factors by Group

Survey respondents indicate that psychological psychosocial factors has the

highest average score on the Likert scale among the four groups of psychosocial factors.

As shown in the table below – Average Likert Scores of Psychosocial Factors by Group,

social relationships and/or living conditions has the next highest average score.

Neighborhood and community psychosocial factors is next, by a very slight margin more

than sociodemographic psychosocial factors.

Average Likert Scores of Psychosocial Factors by Group

Total (n = 164)

Primary Care Physicians

(n = 36)

Nurse Practitioners & Diabetes Educators (n = 128) p value

Psychological 4.54 (.510) 4.18 (.591) 4.64 (.434) .000 Social Relationships / Living Conditions 4.45 (.672) 4.02 (.832)a 4.58 (.566)b .001

Neighborhood/ Community 4.27 (.599) 3.77 (.772)c 4.41 (.458)d .000

Sociodemographic 4.26 (.464) 3.97 (.527) 4.34 (.414) .000 Note. Respondents were asked to indicate the importance of individual psychosocial factors in making, or providing input into, diabetes care clinical decisions. Responses were totaled by group. Responses were captured in a Likert scale: 5 – Very Important, 4 – Important, 3 – Neither Important nor Unimportant, 2 – Unimportant, 1 – Very Unimportant. Standard deviations listed in parenthesis. a n=35. b n = 124. c n = 34. d n = 125.

257

REFERENCES

258

AHIMA. (2014). Data Elements of EHR Documentation. Retrieved June 10, 2014, from

http://library.ahima.org/xpedio/groups/public/documents/ahima/bok1_034460.hcsp?dDocName=bok1_034460

Ahmad, F., Shakya, Y., Li, J., Khoaja, K., Norman, C. D., Lou, W., . . . Ahmadzi, H. M. (2012). A pilot with computer-assisted psychosocial risk-assessment for refugees. BMC Medical Informatics and Decision Making, 12(71), 8. doi: 10.1186/1472-6947-12-71

Ajami, S., & Arab-Chadegani, R. (2013). Barriers to implement Electronic Health Records (EHRs). Materia Sociomedica, 25(3), 213-215.

Al Sayah, F., Williams, B., & Johnson, J. A. (2013). Measuring health literacy in individuals with diabetes: a systematic review and evaluation of available measures. Health Educ Behav, 40(1), 42-55. doi: 10.1177/1090198111436341

American Association of Diabetes Educators. (2015). My AADE Network. Retrieved September 12, 2014, from http://www.diabeteseducator.org/About/membership/myaadenetwork.html

American Diabetes Association. (2011). American Diabetes Association: diagnosis and classification of diabetes mellitus. Diabetes Care, 34(Supplement 1), S62-S69. doi: 10.2337/dc11-S062

American Diabetes Association. (2015). Standards of medical care in diabetes—2015. Diabetes Care, 36(Supplement 1), S11-S66. doi: 10.2337/dc13-S011

Anderson, D. R., & Christison-Lagay, J. (2008). Diabetes self-management in a community health center: improving health behaviors and clinical outcomes for underserved patients. Clinical Diabetes, 26(1), 22-27. doi: 10.2337/diaclin.26.1.22

Anderson, R. J., Freedland, K. E., Clouse, R. E., & Lustman, P. J. (2001). The prevalence of comorbid depression in adults with diabetes: a meta-analysis. Diabetes Care, 24(6), 1069-1078. doi: 10.2337/diacare.24.6.1069

Anderson, R. M., Fitzgerald, J. T., Gorenflo, D. W., & Oh, M. S. (1993). A comparison of the diabetes-related attitudes of health care professionals and patients Patient Educ Couns, 21, 41-50.

Anderson, R. M., & Funnell, M. M. (2005). Patient empowerment: reflections on the challenge of fostering the adoption of a new paradigm. Patient Education and Counseling, 57(2), 153-157. doi: 10.1016/j.pec.2004.05.008

Appiah, B., Hong, Y., Ory, M. G., Helduser, J. W., Begaye, D., Bolin, J. N., & Forjuoh, S. N. (2013). Challenges and opportunities for implementing diabetes self-management guidelines. The Journal of the American Board of Family Medicine, 26(1), 90-92. doi: 10.3122/jabfm.2013.01.120177

Arar, N. H., Wen, L., McGrath, J., Steinbach, R., & Pugh, J. A. (2005). Communicating about medications during primary care outpatient visits: the role of electronic medical records. Inform Prim Care, 13(1), 13-22.

259

Arigo, D., Smyth, J. M., Haggerty, K., & Raggio, G. A. (2014). The social context of the relationship between glycemic control and depressive symptoms in type 2 diabetes. Chronic Illness. doi: 10.1177/1742395314531990

Asan, O., Montague, E. N. H., & Xu, J. (2012). Assessing Patient and Doctor Eye Gaze Patterns Between Two Styles of Doctor Ehr Use in Primary Care Encounters. Proceedings of the Human Factors and Ergonomics Society Annual Meeting, 56(1), 951-955. doi: 10.1177/1071181312561199

Ash, J. S., Sittig, D. F., Dykstra, R. H., Guappone, K., Carpenter, J. D., & Seshadri, V. (2007). Categorizing the unintended sociotechnical consequences of computerized provider order entry. International journal of medical informatics, 76, Supplement 1(0), S21-S27. doi: 10.1016/j.ijmedinf.2006.05.017

Auchincloss, A. H., Roux, A., Mujahid, M. S., Shen, M., Bertoni, A. G., & Carnethon, M. R. (2009). Neighborhood resources for physical activity and healthy foods and incidence of type 2 diabetes mellitus: The multi-ethnic study of atherosclerosis. Archives of Internal Medicine, 169(18), 1698-1704. doi: 10.1001/archinternmed.2009.302

Baker, J. G., & Johnston, M. V. (2000). Prevalence and identification of problems in daily functioning in a primary medicine clinic. Disability and Rehabilitation, 22(16), 716-724. doi: 10.1080/09638280050191981

Barquera, S., Hernandez-Barrera, L., Tolentino, M. L., Espinosa, J., Ng, S. W., Rivera, J. A., & Popkin, B. M. (2008). Energy intake from beverages is increasing among Mexican adolescents and adults. The Journal of Nutrition, 138(12), 2454-2461. doi: 10.3945/jn.108.092163

Bartholomew, D. J., Steele, F., Galbraith, J., & Moustaki, I. (2008). Analysis of Multivariate Social Science Data (2nd ed.).

Bartley, M. (2006). Capabilities and Resilience: Beating the Odds. London. Bass, E. J., DeVoge, J. M., Waggoner-Fountain, L. A., & Borowitz, S. M. (2012).

Resident physicians as human information systems: sources yet seekers. Journal of the American Medical Informatics Association. doi: 10.1136/amiajnl-2012-001112

Beatty, P. C. (2003). Answerable Questions: Advances in the Methodology for Identifying and Resolving Questionnaire Problems in Survey Research. Doctoral Dissertation. The University of Michigan. Dissertation Abstracts International.

Beatty, P. C., & Willis, G. B. (2007). Research Synthesis: The Practice of Cognitive Interviewing. Public Opinion Quarterly, 71(2), 287-311. doi: 10.1093/poq/nfm006

Beaulac, J., Kristjansson, E., & Cummins, S. (2009). A Systematic Review of Food Deserts, 1966-2007. Preventing Chronic Disease, 6(3), A105.

Behforouz, H. L., Drain, P. K., & Rhatigan, J. J. (2014). Rethinking the social history. New England Journal of Medicine, 371(14), 1277-1279. doi: 10.1056/NEJMp1404846

260

Bennett, G. G., McNeill, L. H., Willin, K. Y., Duncan, D. T., Puelo, E., & Emmons, K. M. (2007). Safe to walk? Neighborhood safety and physical activity among public housing residents. PLoS Med, 4(10), e306. doi: 10.1371/journal.pmed.0040306

Berg, M. (1997). Rationalizing Medical Work: Decision-Support Techniques and Medical Practices. Cambridge, Massachusetts: MIT Press.

Berkman, L. F. (1995). The role of social relations in health promotion. Psychosomatic Medicine, 57(3), 245-254.

Berkman, L. F., & Kawachi, I. (2000). Social Epidemiology: Oxford University Press, USA.

Beverly, E. A., Brooks, K. M., Ritholz, M. D., Abrahamson, M. J., & Weinger, K. (2012). Assessment of emotional struggles in type 2 diabetes: patient perspectives. Diabetes Care, 35(8), e62. doi: 10.2337/dc12-0661

Bielamowicz, M. K., Pope, P., & Rice, C. A. (2013). Sustaining a creative community-based diabetes education program: motivating Texans with type 2 diabetes to do well with diabetes control. The Diabetes Educator, 39(1), 119-127. doi: 10.1177/0145721712470605

Blumer, H. (1969). Symbolic interactionism: perspective and method. Englewood Cliffs, N.J: Prentice-Hall.

Bodenheimer, T. (2008). Coordinating care — A perilous journey through the health care system. New England Journal of Medicine, 358(10), 1064-1071. doi: doi:10.1056/NEJMhpr0706165

Bowen, M. E., Cavanaugh, K. L., Wolff, K., Davis, D., Gregory, B., & Rothman, R. L. (2013). Numeracy and dietary intake in patients with type 2 diabetes. The Diabetes Educator, 39(2), 240-247. doi: 10.1177/0145721713475841

Bradburn, N. M., Sudman, S., & Wansink, B. (2004). Asking Questions: The Definitive Guide to Questionnaire Design -- For Market Research, Political Polls, and Social and Health Questionnaires: John Wiley & Sons.

Britten, N. (1995). Qualitative interviews in medical research. British Medical Journal, 311, 251-253.

Brotman, D. J., Golden, S. H., & Wittstein, I. S. (2007). The cardiovascular toll of stress. The Lancet, 370(9592), 1089-1100. doi: 10.1016/s0140-6736(07)61305-1

Bruce, M. A., Beech, B. M., Sims, M., Brown, T. N., Wyatt, S. B., Taylor, H. A., . . . Crook, E. (2009). Social environmental stressors, psychological factors, and kidney disease. Journal of Investigative Medicine, 57(4), 583-589. doi: 10.231/JIM.0b013e31819dbb91

Bruner, E. J., Kivimäki, M., Siegrist, J., Theorell, T., Luukkonen, R., Riihimäki, H., . . . Leino-Arjas, P. (2004). Is the effect of work stress on cardiovascular mortality confounded by socioeconomic factors in the Valmet study? Journal of Epidemiology and Community Health, 58(12), 1019-1020. doi: 10.1136/jech.2003.016881

261

Bryant, A., & Charmaz, K. (2007). The SAGE Handbook of Grounded Theory. London, England: SAGE Publications Ltd.

Burke, H. B., Sessums, L. L., Hoang, A., Becher, D. A., Fontelo, P., Liu, F., . . . Gimbel, R. W. (2015). Electronic health records improve clinical note quality. Journal of the American Medical Informatics Association : JAMIA, 22(1), 199-205. doi: 10.1136/amiajnl-2014-002726

Cadzow, R. B., Vest, B. M., Craig, M., Rowe, J. S., & Kahn, L. S. (2014). “Living Well with Diabetes”: evaluation of a pilot program to promote diabetes prevention and self-management in a medically underserved community. Diabetes Spectrum, 27(4), 246-255. doi: 10.2337/diaspect.27.4.246

Campion Jr., T. R., Waitman, L. R., May, A. K., Ozdas, A., Lorenzi, N. M., & Gadd, C. S. (2010). Social, organizational, and contextual characteristics of clinical decision support systems for intensive insulin therapy: a literature review and case study. International journal of medical informatics, 79(1), 31-43. doi: 10.1016/j.ijmedinf.2009.09.004

Capps, K. H., Mau, S., & Guilford-Blake, R. (2010). Core value, community connections: care coordination in the medical home. Core Value, Community Connections. Retrieved from: http://www.pcpcc.net/sites/default/files/media/carecoordination_pcpcc.pdf

Carey, R. F. (2003). Narrative and information: an ethnography of a lupus support group. (Ph.D.), The University of Western Ontario, London, Ontario, Canada. Retrieved from http://search.proquest.com.proxy.lib.umich.edu/dissertations/docview/305254918/9A979627CA344B00PQ/1?accountid=14667

Cavanaugh, K. L. (2007). Diabetes Management Issues for Patients With Chronic Kidney Disease. Clinical Diabetes, 25(3), 90-97. doi: 10.2337/diaclin.25.3.90

Cefalu, W. T., & Watson, K. (2008). Intensive glycemic control and cardiovascular disease observations from the ACCORD study: now what can a clinician possibly think? Diabetes, 57(5), 1163-1165. doi: 10.2337/db08-0220

Center for Disease Control. (2011). Number (in Millions) of Civilian, Noninstitutionalized Persons with Diagnosed Diabetes, United States, 1980–2011. Diabetes Data & Trends. Retrieved April 21, 2013, from http://www.cdc.gov/diabetes/statistics/prev/national/figpersons.htm

Cerdá, M., Diez-Roux, A. V., Tchetgen, E. T., Gordon-Larsen, P., & Kiefe, C. I. (2010). The relationship between neighborhood poverty and alcohol use: estimation by marginal structural models. Epidemiology, 21(4), 482-489. doi: 10.1097/EDE.0b013e3181e13539

Ceriello, A., Barkai, L., Christiansen, J. S., Czupryniak, L., Gomis, R., Harno, K., . . . Wens, J. (2012). Diabetes as a case study of chronic disease management with a personalized approach: The role of a structured feedback loop. Diabetes Research and Clinical Practice, 98(1), 5-10. doi: 10.1016/j.diabres.2012.07.005

262

Charmaz, K. (2006). Constructing grounded theory: a practical guide through qualitative analysis. London: SAGE.

Chen, I. J., & Paulraj, A. (2004). Towards a theory of supply chain management: the constructs and measurements. Journal of Operations Management, 22(2), 119-150. doi: 10.1016/j.jom.2003.12.007

Chida, Y., & Hamer, M. (2008). An association of adverse psychosocial factors with diabetes mellitus: a meta-analytic review of longitudinal cohort studies. Diabetologia, 51(12), 2168-2178. doi: 10.1007/s00125-008-1154-1

Cho, Y. I., Johnson, T. P., & VanGeest, J. B. (2013). Enhancing surveys of health care professionals: a meta-analysis of techniques to improve response. Evaluation & the Health Professions, 36(3), 382-407. doi: 10.1177/0163278713496425

Chunchu, K., Mauksch, L., Charles, C., Ross, V., & Pauwels, J. (2012). A patient centered care plan in the EHR: Improving collaboration and engagement. Families, Systems, & Health, 30(3), 199-209. doi: 10.1037/a0029100

Ciechanowski, P. S., Katon, W. J., Russo, J. E., & Walker, E. A. (2001). The patient-provider relationship: Attachment Theory and adherence to treatment in diabetes. American Journal of Psychiatry, 158(1), 29-35. doi: 10.1176/appi.ajp.158.1.29

Clark, C. R., Ommerborn, M. J., Hickson, D. A., Grooms, K. N., Sims, M., Taylor, H. A., & Albert, M. A. (2013). Neighborhood disadvantage, neighborhood safety and cardiometabolic risk factors in African Americans: biosocial associations in the Jackson Heart study. PLoS ONE, 8(5), e63254. doi: 10.1371/journal.pone.0063254

Clark, L., Vincent, D., Zimmer, L., & Sanchez, J. (2009). Cultural values and political economic contexts of diabetes among low-income Mexican Americans. Journal of Transcultural Nursing, 20(4), 382-394. doi: 10.1177/1043659609334851

Claydon-Platt, K., Manias, E., & Dunning, T. (2014). The barriers and facilitators people with diabetes from a nonEnglish speaking background experience when managing their medications: a qualitative study. Journal of Clinical Nursing, 23(15-16), 2234-2246. doi: 10.1111/jocn.12501

Cohen, D. A., Mason, K., Bedimo, A., Scribner, R., Basolo, V., & Farley, T. A. (2003). Neighborhood physical conditions and health. American Journal of Public Health, 93(3), 467-471.

Cohen, J. (1988). Statistical Power Analysis for the Behavioral Sciences (2nd ed.): Lawrence Erlbaum Associates.

Colbert, C. Y., Diaz-Guzman, E., Myers, J. D., & Arroliga, A. C. (2013). How to interpret surveys in medical research: A practical approach. Cleveland Clinic Journal of Medicine, 80(7), 423-425, 430-432, 434-435. doi: 10.3949/ccjm.80a.12122

Cook, D. A., & Beckman, T. J. (2006). Current concepts in validity and reliability for psychometric instruments: theory and application. The American Journal of Medicine, 119(2), 166.e167-116. doi: 10.1016/j.amjmed.2005.10.036

263

Cook, D. J., Greengold, N. L., Ellrodt, A. G., & Weingarten, S. R. (1997). The relation between systematic reviews and practice guidelines. Annals of Internal Medicine, 127(3), 210-216. doi: 10.7326/0003-4819-127-3-199708010-00006

Corbin, J., & Strauss, A. (2007). Basics of Qualitative Research: Techniques and Procedures for Developing Grounded Theory (3rd ed.): SAGE Publications.

Cosansu, G., & Erdogan, S. (2014). Influence of psychosocial factors on self-care behaviors and glycemic control in Turkish patients with type 2 diabetes mellitus. Journal of Transcultural Nursing, 25(1), 51-59. doi: 10.1177/1043659613504112

Cossrow, N., & Falkner, B. (2004). Race/ethnic issues in obesity and obesity-related comorbidities. The Journal of Clinical Endocrinology & Metabolism, 89(6), 2590-2594. doi: 10.1210/jc.2004-0339

Covell, D. G., Uman, G. C., & Manning, P. R. (1985). Information needs in office practice: are they being met? Annals of Internal Medicine, 103(4), 596-599. doi: 10.7326/0003-4819-103-4-596

Creswell, J. W. (2006). Qualitative Inquiry and Research Design: Choosing Among Five Approaches (2nd ed.): SAGE Publications.

Creswell, J. W. (2013). Research Design: Qualitative, Quantitative, and Mixed Methods Approaches (4th ed.): SAGE Publications.

Curry, L. A., Nembhard, I. M., & Bradley, E. H. (2009). Qualitative and Mixed Methods Provide Unique Contributions to Outcomes Research. Circulation, 119(10), 1442-1452. doi: 10.1161/circulationaha.107.742775

Custers, E. J. F. M. (2013). Medical education and cognitive continuum theory: an alternative perspective on medical problem solving and clinical reasoning. Academic Medicine, 88(8), 1074-1080. doi: 10.1097/ACM.0b013e31829a3b10

Dahmann, N., Wolch, J., Joassart-Marcelli, P., Reynolds, K., & Jerrett, M. (2010). The active city? Disparities in provision of urban public recreation resources. Health & Place, 16(3), 431-445. doi: 10.1016/j.healthplace.2009.11.005

Daniulaityte, R. (2004). Making sense of diabetes: cultural models, gender and individual adjustment to Type 2 diabetes in a Mexican community. Social Science & Medicine, 59(9), 1899-1912. doi: 10.1016/j.socscimed.2004.03.001

Davidson, M. B. (2007). The effectiveness of nurse- and pharmacist-directed care in diabetes disease management: a narrative review. Current Diabetes Reviews, 3(4), 280-286. doi: 10.2174/157339907782330058

Davidson, M. B. (2009). How our current medical care system fails people with diabetes: lack of timely, appropriate clinical decisions. Diabetes Care, 32(2), 370-372. doi: 10.2337/dc08-2046

de Alba Garcia, J. G., Rocha, A. L. S., Lopez, I., Baer, R. D., Dressler, W., & Weller, S. C. (2007). “Diabetes is my companion”: Lifestyle and self-management among good and poor control Mexican diabetic patients. Social Science & Medicine, 64(11), 2223-2235. doi: 10.1016/j.socscimed.2007.02.001

264

De Faoite, D., & Hanson, B. (2015). Multicultural aspects affecting patient-doctor interactions and the conduct of clinical research examples drawn from the literature and first-hand knowledge. GSTF Journal of Nursing and Health Care, 2(1). doi: 10.5176/2345-718x_2.1.64

Del Fiol, G., Workman, T. E., & Gorman, P. N. (2014). Clinical questions raised by clinicians at the point of care: a systematic review. JAMA Intern Med, 174(5), 710-718. doi: 10.1001/jamainternmed.2014.368

Delamater, A. M. (2006). Improving patient adherence. Clinical Diabetes, 24(2), 71-77. doi: 10.2337/diaclin.24.2.71

Delamater, A. M., Jacobson, A. M., Anderson, B., Cox, D., Fisher, L., Lustman, P., . . . Wysocki, T. (2001). Psychosocial therapies in diabetes: report of the psychosocial therapies working group. Diabetes Care, 24(7), 1286-1292. doi: 10.2337/diacare.24.7.1286

Di lorio, C. K. (2006). Measurement in Health Behavior: Methods for Research and Evaluation (Vol. Volume 1 of Jossey-Bass Public Health): Wiley.

Donaldson, M. S., Yordy, K. D., Lohr, K. N., & Vanselow, N. A. (1996). Primary Care: America's Health in a New Era: National Academies Press.

Dowding, D. (2008). Computerized decision support systems in nursing. In N. Cullum, D. Ciliska, B. Haynes, & S. Marks (Eds.), Evidence-based Nursing: An Introduction (pp. 271-276): Wiley.

Doyle, R. J., Wang, N., Anthony, D., Borkan, J., Shield, R. R., & Goldman, R. E. (2012). Computers in the examination room and the electronic health record: physicians' perceived impact on clinical encounters before and after full installation and implementation. Family Practice, 29(5), 601-608. doi: 10.1093/fampra/cms015

Drennan, J. (2003). Cognitive interviewing: verbal data in the design and pretesting of questionnaires. Journal of Advanced Nursing, 42(1), 57-63. doi: 10.1046/j.1365-2648.2003.02579.x

Dreyfus, H. L., Dreyfus, S. E., & Athanasiou, T. (1986). Mind Over Machine: The Power of Human Intuition and Expertise in the Era of the Computer: Free Press.

Dubois, R. W., Chawla, A. J., Neslusan, C. A., Smith, M. W., & Wade, S. (2000). Explaining drug spending trends: does perception match reality? Health affairs, 19(2), 231-239. doi: 10.1377/hlthaff.19.2.231

DuBose, T. D., Behrens, M. T., Berns, A., Davis, C., Goldfarb, S., Hostetter, T., . . . Himmelfarb, J. (2009). The Patient-Centered Medical Home and Nephrology. Journal of the American Society of Nephrology, 20(4), 681-682. doi: 10.1681/asn.2009010012

Duke, S.-A., Colagiuri, S., & Colagiuri, R. (2009). Individual patient education for people with type 2 diabetes mellitus (Review). The Cochrane Library, (1). Retrieved from: http://onlinelibrary.wiley.com/doi/10.1002/14651858.CD005268.pub2/pdf/standard doi:10.1002/14651858.CD005268.pub2

265

Egan, M., Tannahill, C., Petticrew, M., & Thomas, S. (2008). Psychosocial risk factors in home and community settings and their associations with population health and health inequalities: A systematic meta-review. BMC Public Health, 8(1), 239. doi: 10.1186/1471-2458-8-239

Egginton, J., Ridgeway, J., Shah, N., Balasubramaniam, S., Emmanuel, J., Prokop, L., . . . Murad, M. (2012). Care management for type 2 diabetes in the United States: a systematic review and meta-analysis. BMC Health Services Research, 12(1), 72.

Eldor, R., & Raz, I. (2009). American Diabetes Association indications for statins in diabetes: is there evidence? Diabetes Care, 32(suppl 2), S384-S391. doi: 10.2337/dc09-S345

Ellis, S. E., Speroff, T., Dittus, R. S., Brown, A., Pichert, J. W., & Elasy, T. A. (2004). Diabetes patient education: a meta-analysis and meta-regression. Patient Education and Counseling, 52(1), 97-105. doi: 10.1016/S0738-3991(03)00016-8

Esposito, K., Chiodini, P., Bellastella, G., Maiorino, M. I., & Giugliano, D. (2012). Proportion of patients at HbA1c target <7% with eight classes of antidiabetic drugs in type 2 diabetes: systematic review of 218 randomized controlled trials with 78 945 patients. Diabetes, Obesity and Metabolism, 14(3), 228-233. doi: 10.1111/j.1463-1326.2011.01512.x

Estabrooks, P. A., Boyle, M., Emmons, K. M., Glasgow, R. E., Hesse, B. W., Kaplan, R. M., . . . Taylor, M. V. (2012). Harmonized patient-reported data elements in the electronic health record: supporting meaningful use by primary care action on health behaviors and key psychosocial factors. Journal of the American Medical Informatics Association, 19(4), 575-582. doi: 10.1136/amiajnl-2011-000576

Evans, L., & Trotter, D. R. M. (2009). Epistemology and uncertainty in primary care: an exploratory study. Fam Med, 41(5), 319-326.

Evert, A. B., Boucher, J. L., Cypress, M., Dunbar, S. A., Franz, M. J., Mayer-Davis, E. J., . . . Yancy, W. S. (2014). Nutrition therapy recommendations for the management of adults with diabetes. Diabetes Care, 37(Supplement 1), S120-S143. doi: 10.2337/dc14-S120

Fetters, M. D., Curry, L. A., & Creswell, J. W. (2013). Achieving integration in mixed methods designs-principles and practices. Health Services Research, 48(6 Pt 2), 2134-2156. doi: 10.1111/1475-6773.12117

Fisher, L., Mullan, J. T., Arean, P., Glasgow, R. E., Hessler, D., & Masharani, U. (2010). Diabetes distress but not clinical depression or depressive symptoms Is associated with glycemic control in both cross-sectional and longitudinal analyses. Diabetes Care, 33(1), 23-28. doi: 10.2337/dc09-1238

Fonseca, V., & Clark, N. G. (2006). Standards of medical care in diabetes: response to power. Diabetes Care, 29(2), 476-477. doi: 10.2337/diacare.29.02.06.dc05-1989

Friedman, C. P., & Wyatt, J. C. (2006). Evaluation Methods in Biomedical Informatics: Springer.

266

Frølich, A., Bellows, J., Nielsen, B., Brockhoff, P. B., & Hefford, M. (2010). Effective population management practices in diabetes care - an observational study. BMC Health Services Research, 10(1), 277.

Funnell, M. M. (2006). The Diabetes Attitudes, Wishes, and Needs (DAWN) Study. Clinical Diabetes, 24(4), 154-155. doi: 10.2337/diaclin.24.4.154

Funnell, M. M. (2010). Peer-based behavioural strategies to improve chronic disease self-management and clinical outcomes: evidence, logistics, evaluation considerations and needs for future research. Family Practice, 27(suppl 1), i17-i22. doi: 10.1093/fampra/cmp027

Funnell, M. M., Brown, T. L., Childs, B. P., Haas, L. B., Hosey, G. M., Jensen, B., . . . Weiss, M. A. (2009). National standards for diabetes self-management education. Diabetes Care, 32(Supplement 1), S87-S94. doi: 10.2337/dc09-S087

Gandara, B. K., & Morton, T. H. (2011). Non-periodontal oral manifestations of diabetes: a framework for medical care providers. Diabetes Spectrum, 24(4), 199-205. doi: 10.2337/diaspect.24.4.199

Ganz, P. (2008). Institute of Medicine report: recognizing psychosocial health needs to treat the whole patient. Journal of Oncology Practice, 4(3), 128-130. doi: 10.1200/jop.0833003

Gaucher, N., Lantos, J., & Payot, A. (2013). How do national guidelines frame clinical ethics practice? A comparative analysis of guidelines from the US, the UK, Canada and France. Social Science & Medicine, 85, 74-78. doi: 10.1016/j.socscimed.2013.02.038

Genesis Physicians Group. (2015). About Us. Retrieved September 10, 2014, from www.genesisdocs.org

Gigante, J. (2013). Teaching Clinical Reasoning Skills to Help your Learners ?Get? the Diagnosis. Pediatrics & Therapeutics, 03(04). doi: 10.4172/2161-0665.1000e121

Giugliano, D., Maiorino, M. I., Bellastella, G., Chiodini, P., Ceriello, A., & Esposito, K. (2011). Efficacy of insulin analogs in achieving the hemoglobin A1c target of <7% in type 2 diabetes: meta-analysis of randomized controlled trials. Diabetes Care, 34(2), 510-517. doi: 10.2337/dc10-1710

Glaser, B. G. (1978). Theoretical sensitivity: advances in the methodology of grounded theory: Sociology Press.

Glasgow, R. E., Kaplan, R. M., Ockene, J. K., Fisher, E. B., & Emmons, K. M. (2012). Patient-reported measures of psychosocial issues and health behavior should be added to electronic health records. Health affairs, 31(3), 497-504. doi: 10.1377/hlthaff.2010.1295

Goebel, S., & Mehdorn, H. M. (2011). Measurement of psychological distress in patients with intracranial tumours: the NCCN distress thermometer. J Neurooncol, 104(1), 357-364. doi: 10.1007/s11060-010-0501-5

Goldstein, M. K., Lavori, P., Coleman, R., Advani, A., & Hoffman, B. B. (2005). Improving adherence to guidelines for hypertension drug prescribing: cluster-

267

randomized controlled trial of general versus patient-specific recommendations. AJMC - American Journal of Managed Care, 11(11), 677-685.

Gorman, P. N., & Helfand, M. (1995). Information Seeking in Primary Care: How Physicians Choose Which Clinical Questions to Pursue and Which to Leave Unanswered. Medical Decision Making, 15(2), 113-119. doi: 10.1177/0272989x9501500203

Grava-Gubins, I., & Scott, S. (2008). Effects of various methodologic strategies: Survey response rates among Canadian physicians and physicians-in-training. Canadian Family Physician, 54(10), 1424-1430.

Green, E. P., Wendland, J., Carver, M. C., Hughes Rinker, C., & Mun, S. K. (2012). Lessons learned from implementing the patient-centered medical home. International Journal of Telemedicine and Applications, 2012, 8. doi: 10.1155/2012/103685

Gross, R., Tabenkin, H., Porath, A., Heymann, A., Greenstein, M., Porter, B., & Matzliach, R. (2003). The relationship between primary care physicians’ adherence to guidelines for the treatment of diabetes and patient satisfaction: findings from a pilot study. Family Practice, 20(5), 563-569. doi: 10.1093/fampra/cmg512

Groves, R. M., Fowler, F. J., Couper, M. P., Lepkowski, J. M., Singer, E., & Tourangeau, R. (2011). Survey Methodology: John Wiley & Sons.

Guest, G., Bunce, A., & Johnson, L. (2006). How Many Interviews Are Enough? An Experiment with Data Saturation and Variability. Field Methods, 18(1), 59-82. doi: 10.1177/1525822x05279903

Haase, A., Follmann, M., Skipka, G., & Kirchner, H. (2007). Developing search strategies for clinical practice guidelines in SUMSearch and Google Scholar and assessing their retrieval performance. BMC Medical Research Methodology, 7(28). doi: 10.1186/1471-2288-7-28

Haggerty, J. L., Roberge, D., Freeman, G. K., Beaulieu, C., & Bréton, M. (2012). Validation of a generic measure of continuity of care: when patients encounter several clinicians. The Annals of Family Medicine, 10(5), 443-451. doi: 10.1370/afm.1378

Hall, A., & Walton, G. (2004). Information overload within the health care system: a literature review. Health Information & Libraries Journal, 21(2), 102-108. doi: 10.1111/j.1471-1842.2004.00506.x

Hall, P. A., & Taylor, R. C. (2009). Health, Social Relations, and Public Policy. In P. A. Hall & M. Lamont (Eds.), Successful Societies How Institutions and Culture Affect Health (pp. 82-103): Cambridge University Press.

Halladay, J. R., DeWalt, D. A., Wise, A., Qaqish, B., Reiter, K., Lee, S.-Y., . . . Donahue, K. E. (2014). More extensive implementation of the chronic care model is associated with better lipid control in diabetes. The Journal of the American Board of Family Medicine, 27(1), 34-41. doi: 10.3122/jabfm.2014.01.130070

268

Halonen, J. I., Kivimäki, M., Pentti, J., Kawachi, I., Virtanen, M., Martikainen, P., . . . Vahtera, J. (2012). Quantifying neighbourhood socioeconomic effects in clustering of behaviour-related risk factors: a multilevel analysis. PLoS ONE, 7(3), e32937. doi: 10.1371/journal.pone.0032937

Hamm, R. M. (1988). Clinical intuition and clinical analysis: expertise and the cognitive continuum. In J. Dowie & A. S. Elstein (Eds.), Professional Judgment: A Reader in Clinical Decision Making (pp. 78-106): Cambridge University Press.

Harris, L. R., & Brown, G. T. L. (2010). Mixing Interview and questionnaire methods-Practical problems in aligning data. Practical Assessment, Research & Evaluation, 15(1).

Haskard, K. B., Williams, S. L., & DiMatteo, M. R. (2009). Physician-patient communication: psychosocial care, emotional-well being and health outcomes. In D. E. Brashers & D. J. Goldsmith (Eds.), Communicating to Manage Health and Illness: Routledge/Taylor and Francis Group.

Hawley, K. M., Cook, J. R., & Jensen-Doss, A. (2009). Do noncontingent incentives increase survey response rates among mental health providers? A randomized trial comparison. Adm Policy Ment Health, 36(5), 343-348. doi: 10.1007/s10488-009-0225-z

Heisler, M., & Resnicow, K. (2008). Helping patients make and sustain healthy changes: a brief introduction to motivational interviewing in clinical diabetes care. Clinical Diabetes, 26(4), 161-165. doi: 10.2337/diaclin.26.4.161

Henry, S. L., Rook, K. S., Stephens, M. A. P., & Franks, M. M. (2013). Spousal undermining of older diabetic patients’ disease management. Journal of Health Psychology, 18(12), 1550-1561. doi: 10.1177/1359105312465913

Hertzman, C., & Siddiqi, A. (2009). Population Health and the Dynamics of Collective Development. In P. A. Hall & M. Lamont (Eds.), Successful Societies How Institutions and Culture Affect Health (pp. 23-52): Cambridge University Press.

Hesse-Biber, S. N., & Leavy, P. (2010). The Practice of Qualitative Research: SAGE Publications.

Higgs, J. (2008). Clinical reasoning in the health professions (3rd ed.). Amsterdam ; Boston: Elsevier Health Sciences.

Hillestad, R., Bigelow, J., Bower, A., Girosi, F., Meili, R., Scoville, R., & Taylor, R. (2005). Can Electronic Medical Record Systems Transform Health Care? Potential Health Benefits, Savings, And Costs. Health affairs, 24(5), 1103-1117. doi: 10.1377/hlthaff.24.5.1103

Hirsch, I. B., Amiel, S. A., Blumer, I. R., Bode, B. W., Edelman, S. V., Seley, J. J., . . . Kilpatrick, E. S. (2012). Using multiple measures of glycemia to support individualized diabetes management: recommendations for clinicians, patients, and payers. Diabetes Technol Ther, 14(11), 973-983; quiz 983. doi: 10.1089/dia.2012.0132

269

Hogg, W., & Dyke, E. (2011). Improving measurement of primary care system performance. Canadian Family Physician, 57(7), 758-760.

Holbrook, A., Thabane, L., Keshavjee, K., Dolovich, L., Bernstein, B., Chan, D., . . . Gerstein, H. (2009). Individualized electronic decision support and reminders to improve diabetes care in the community: COMPETE II randomized trial. Canadian Medical Association Journal, 181(1-2), 37-44. doi: 10.1503/cmaj.081272

Holland, J. C., & Bultz, B. D. (2007). The NCCN guideline for distress management: a case for making distress the sixth vital sign. Journal of the National Comprehensive Cancer Network, 5(1), 3-7.

Holm-Denoma, J., Smith, A., Lewinsohn, P., & Pettit, J. (2014). Psychosocial predictors of body mass index at late childhood: A longitudinal investigation. Journal of Health Psychology, 19(6), 754-764. doi: 10.1177/1359105313479626

Huang, E. S., Brown, S. E. S., Ewigman, B. G., Foley, E. C., & Meltzer, D. O. (2007). Patient perceptions of quality of life with diabetes-related complications and treatments. Diabetes Care, 30(10), 2478-2483. doi: 10.2337/dc07-0499

Hübner-Bloder, G., Duftschmid, G., Kohler, M., Rinner, C., Saboor, S., & Ammenwerth, E. (2011). Clinical situations and information needs of physicians during treatment of diabetes mellitus patients: a triangulation study. Studies in Health Technology and Informatics, 169, 369-373.

Hunt, L. M., Valenzuela, M. A., & Pugh, J. A. (1998). Porque me tocó a mi? Mexican American diabetes patients' causal stories and their relationship to treatment behaviors. Social Science & Medicine, 46(8), 959-969. doi: 10.1016/S0277-9536(97)10014-4

Iida, M., Parris Stephens, M. A., Rook, K. S., Franks, M. M., & Salem, J. K. (2010). When the going gets tough, does support get going? Determinants of spousal support provision to type 2 diabetic patients. Personality and Social Psychology Bulletin, 36(6), 780-791. doi: 10.1177/0146167210369897

Institute of Education Sciences. (2014). Definition of Literacy. National Assessment of Adult Literacy. Retrieved April 3, 2014, from http://nces.ed.gov/naal/fr_definition.asp

Institute of Medicine. (2008a). Cancer Care for the Whole Patient: Meeting Psychosocial Health Needs. Washington, DC: National Academies Press.

Institute of Medicine. (2008b). Institute of Medicine report: recognizing psychosocial health needs to treat the whole patient. Journal of Oncology Practice, 4(3), 128-130. doi: 10.1200/jop.0833003

Institute of Medicine. (2014a). Capturing social and behavioral domains and measures in electronic health records: phase 2. Retrieved from The Institute of Medicine website: http://www.iom.edu/Reports/2014/EHRdomains2.aspx

270

Institute of Medicine. (2014b). Capturing social and behavioral domains in electronic health records: phase 1. Retrieved from: http://iom.edu/Reports/2014/Capturing-Social-and-Behavioral-Domains-in-Electronic-Health-Records-Phase-1.aspx

Institute of Medicine of the National Academies. (2014). Health Literacy and Numeracy - Workshop Summary. Retrieved from: http://www.iom.edu/Reports/2014/Health-Literacy-and-Numeracy.aspx

Irani, J. S., Middleton, J. L., Marfatia, R., Omana, E. T., & D'Amico, F. (2009). The use of electronic health records in the exam room and patient satisfaction: a systematic review. The Journal of the American Board of Family Medicine, 22(5), 553-562. doi: 10.3122/jabfm.2009.05.080259

Jack, L., Jack, N. H., & Hayes, S. C. (2012). Social determinants of health in minority populations: a call for multidisciplinary approaches to eliminate diabetes-related health disparities. Diabetes Spectrum, 25(1), 9-13. doi: 10.2337/diaspect.25.1.9

Jack, L., Liburd, L. C., Tucker, P., & Cockrell, T. (2014). Having their say: patients’ perspectives and the clinical management of diabetes*. Clin Ther, 36(4), 469-476. doi: 10.1016/j.clinthera.2014.02.003

Jansen, A. C., van Aalst-Cohen, E. S., Hutten, B. A., Buller, H. R., Kastelein, J. J., & Prins, M. H. (2005). Guidelines were developed for data collection from medical records for use in retrospective analyses. Journal of clinical epidemiology, 58(3), 269-274.

Johns Hopkins Center to Eliminate Cardiovascular Health Disparities. (2014a). Stable Housing. Retrieved September 6, 2014, from http://www.jhsph.edu/research/centers-and-institutes/johns-hopkins-center-to-eliminate-cardiovascular-health-disparities/about/influences_on_health/stable_housing.html

Johns Hopkins Center to Eliminate Cardiovascular Health Disparities. (2014b). Stable Housing. Retrieved April 7, 2014, from http://www.jhsph.edu/research/centers-and-institutes/johns-hopkins-center-to-eliminate-cardiovascular-health-disparities/about/influences_on_health/stable_housing.html

Jotkowitz, A. B., Rabinowitz, G., Segal, A. R., Weitzman, R., Epstein, L., & Porath, A. (2006). Do patients with diabetes and low socioeconomic status receive less care and have worse outcomes? A national study. The American Journal of Medicine, 119(8), 665-669. doi: 10.1016/j.amjmed.2006.02.010

Kahane, S., Stutz, E., & Aliarzadeh, B. (2011). Must we appear to be all-knowing?: Patients’ and family physicians’ perspectives on information seeking during consultations. Canadian Family Physician, 57(6), e228-e236.

Kahn, R., & Anderson, J. E. (2009). Improving diabetes care: The model for health care reform. Diabetes Care, 32(6), 1115-1118. doi: 10.2337/dc09-0184

Kalyani, R. R., Saudek, C. D., Brancati, F. L., & Selvin, E. (2010). Association of diabetes, comorbidities, and A1C with functional disability in older adults: results from the National Health and Nutrition Examination Survey (NHANES), 1999–2006. Diabetes Care, 33(5), 1055-1060. doi: 10.2337/dc09-1597

271

Kangovi, S., Barg, F. K., Carter, T., Long, J. A., Shannon, R., & Grande, D. (2013). Understanding why patients of low socioeconomic status prefer hospitals over ambulatory care. Health affairs, 32(7), 1196-1203. doi: 10.1377/hlthaff.2012.0825

Karazsia, B. T., Berlin, K. S., Armstrong, B., Janicke, D. M., & Darling, K. E. (2014). Integrating mediation and moderation to advance theory development and testing. Journal of Pediatric Psychology, 39(2), 163-173. doi: 10.1093/jpepsy/jst080

Kassirer, J. P., Wong, J. B., & Kopelman, R. I. (2010). Learning clinical reasoning. Baltimore, MD: Lippincott Williams & Wilkins.

Kay, S., & Purves, I. N. (1996). Medical records and other stories: a narratological framework. Methods of Information in Medicine, 35(2), 72-87.

Kelley, J. M., Kraft-Todd, G., Schapira, L., Kossowsky, J., & Riess, H. (2014). The influence of the patient-clinician relationship on healthcare outcomes: a systematic review and meta-analysis of randomized controlled trials. PLoS ONE, 9(4), e94207. doi: 10.1371/journal.pone.0094207

Kelley, K., Clark, B., Brown, V., & Sitzia, J. (2003). Good practice in the conduct and reporting of survey research. International Journal for Quality in Health Care, 15(3), 261-266. doi: 10.1093/intqhc/mzg031

Kemp, S. P., & Brandwein, R. (2010). Feminisms and social work in the United States: an intertwined history. Affilia: Journal of Women and Social Work, 25(4), 341-364. doi: 10.1177/0886109910384075

Kirpitch, A. R., & Maryniuk, M. D. (2011). The 3 R's of glycemic index: recommendations, research, and the real world. Clinical Diabetes, 29(4), 155-159. doi: 10.2337/diaclin.29.4.155

Kittler, P. G. (1995). Food and Culture in America: A Nutrition Handbook: West Publishing Company, College & School Division.

Kittler, P. G., & Sucher, K. (1995a). Food Culture America: West Publishing Company, College & School Division.

Kittler, P. G., & Sucher, K. (1995b). Food Culture America: West Publishing Company, College & School Division.

Klabunde, C. N., Willis, G. B., & Casalino, L. P. (2013). Facilitators and barriers to survey participation by physicians: a call to action for researchers. Evaluation & the Health Professions, 36(3), 279-295. doi: 10.1177/0163278713496426

Klinkman, M., & van Weel, C. (2011). Prospects for person-centered diagnosis in general medicine. J Eval Clin Pract, 17(2), 365-370. doi: 10.1111/j.1365-2753.2010.01582.x

Kripalani, S., LeFevre, F., Phillips, C. O., Williams, M. V., Basaviah, P., & Baker, D. W. (2007). Deficits in communication and information transfer between hospital-based and primary care physicians: implications for patient safety and continuity of care. Journal of the American Medical Association, 297(8), 831-841. doi: 10.1001/jama.297.8.831

272

Krishnan, S., Cozier, Y. C., Rosenberg, L., & Palmer, J. R. (2010). Socioeconomic status and incidence of type 2 diabetes: results from the Black Women's Health Study. American Journal of Epidemiology, 171(5), 564-570. doi: 10.1093/aje/kwp443

Krist, A. H., Beasley, J. W., Crosson, J. C., Kibbe, D. C., Klinkman, M. S., Lehmann, C. U., . . . Waldren, S. E. (2014a). Electronic health record functionality needed to better support primary care. Journal of the American Medical Informatics Association. doi: 10.1136/amiajnl-2013-002229

Krist, A. H., Beasley, J. W., Crosson, J. C., Kibbe, D. C., Klinkman, M. S., Lehmann, C. U., . . . Waldren, S. E. (2014b). Electronic health record functionality needed to better support primary care. Journal of the American Medical Informatics Association, 21(5), 764-771. doi: 10.1136/amiajnl-2013-002229

Krug, E. G. (2002). World report on violence and health: World Health Organization. Kulkarni, K. D. (2004). Food, culture, and diabetes in the United States. Clinical

Diabetes, 22(4), 190-192. doi: 10.2337/diaclin.22.4.190 LaPiere, R. T. (1934). Attitudes Vs. Actions: University of North Carolina Press. LeCompte, M. D., & Goetz, J. P. (1982). Problems of reliability and validity in

ethnographic research. Review of Educational Research, 52(1), 31-60. doi: 10.3102/00346543052001031

Lelievre, S., & Schultz, K. (2010). Does computer use in patient-physician encounters influence patient satisfaction? Canadian Family Physician, 56(1), e6-e12.

Levinson, W., & Roter, D. (1995). Physicians’ psychosocial beliefs correlate with their patient communication skills. Journal of General Internal Medicine, 10(7), 375-379. doi: 10.1007/BF02599834

Liebow, E. (1967). Tally's Corner: A Study of Negro Streetcorner Men. With a Foreword by Hylan Lewis. Boston: Little, Brown.

Lillrank, A. (2012). Managing the Interviewer Self. In J. F. Gubrium (Ed.), The SAGE Handbook of Interview Research: The Complexity of the Craft (pp. 281-294): SAGE Publications.

Little, R. J. A., & Rubin, D. B. (2002). Statistical Analysis with Missing Data (2nd ed.): Wiley.

Lohr, K. N. (1994). Guidelines for clinical practice: applications for primary care. International Journal for Quality in Health Care, 6(1), 17-25. doi: 10.1093/intqhc/6.1.17

Lomas, J., & Haynes, R. B. (1988). A taxonomy and critical review of tested strategies for the application of clinical practice recommendations: from "official" to "individual" clinical policy. American Journal of Preventive Medicine, 4(4 Suppl), 77-94; discussion 95-77.

Loveman, E., Cave, C., Green, C., Royle, P., Dunn, N., & Waugh, N. (2003). The clinical and cost-effectiveness of patient education models for diabetes: a systematic review and economic evaluation. Health Technology Assessment, 7(22).

273

Macleod, J., & Davey Smith, G. (2003). Psychosocial factors and public health: a suitable case for treatment? Journal of Epidemiology and Community Health, 57(8), 565-570. doi: 10.1136/jech.57.8.565

MacMullin, S. E., & Taylor, R. S. (1984). Problem dimensions and information traits. The Information Society, 3(1), 91-111. doi: 10.1080/01972243.1984.9959994

Makoul, G., Curry, R. H., & Tang, P. C. (2001). The use of electronic medical records: communication patterns in outpatient encounters. Journal of the American Medical Informatics Association, 8(6), 610-615. doi: 10.1136/jamia.2001.0080610

Margalit, R. S., Roter, D., Dunevant, M. A., Larson, S., & Reis, S. (2006). Electronic medical record use and physician–patient communication: An observational study of Israeli primary care encounters. Patient Education and Counseling, 61(1), 134-141. doi: 10.1016/j.pec.2005.03.004

Marshall, M. N. (1996). Sampling for qualitative research. Family Practice, 13(6), 522-526. doi: 10.1093/fampra/13.6.522

Martikainen, P., Bartley, M., & Lahelma, E. (2002). Psychosocial determinants of health in social epidemiology. International Journal of Epidemiology, 31(6), 1091-1093. doi: 10.1093/ije/31.6.1091

Mayberry, L. S., & Osborn, C. Y. (2012). Family support, medication adherence, and glycemic control among adults with type 2 diabetes. Diabetes Care, 35(6), 1239-1245. doi: 10.2337/dc11-2103

McEwen, B. S. (1998). Protective and damaging effects of stress mediators. New England Journal of Medicine, 338(3), 171-179. doi: doi:10.1056/NEJM199801153380307

McGlynn, E. A., Asch, S. M., Adams, J., Keesey, J., Hicks, J., DeCristofaro, A., & Kerr, E. A. (2003). The quality of health care delivered to adults in the United States. New England Journal of Medicine, 348(26), 2635-2645. doi: 10.1056/NEJMsa022615

McKinlay, J. B., Piccolo, R., & Marceau, L. (2013). An additional cause of health care disparities: the variable clinical decisions of primary care doctors. J Eval Clin Pract, 19(4), 664-673. doi: 10.1111/jep.12015

Meddings, J., Kerr, E. A., Heisler, M., & Hofer, T. P. (2012). Physician assessments of medication adherence and decisions to intensify medications for patients with uncontrolled blood pressure: still no better than a coin toss. BMC Health Services Research, 12(1), 270.

Miles, M. B., Huberman, A. M., & Saldaña, J. (2014). Qualitative Data Analysis: A Methods Sourcebook. Thousand Oaks: Sage.

Mirowsky, J., & Ross, C. E. (1998). Education, personal control, lifestyle and health: a human capital hypothesis. Research on Aging, 20(4), 415-449. doi: 10.1177/0164027598204003

274

Monajemi, A., Rostami, E. R., Savaj, S., & Rikers, R. M. J. P. (2012). How does patient management knowledge integrate into an illness script? Education for health, 25(3), 153-159. doi: 10.4103/1357-6283.109791

Monajemi, A., Schmidt, H. G., & Rikers, R. M. J. P. (2012). Assessing patient management plans of doctors and medical students: an illness script perspective. The Journal of continuing education in the health professions, 32(1), 4-9. doi: 10.1002/chp.21118

Morris, A. H. (2000). Developing and implementing computerized protocols for standardization of clinical decisions. Annals of Internal Medicine, 132(5), 373-383. doi: 10.7326/0003-4819-132-5-200003070-00007

Morse, J. M. (2010). Sampling in Grounded Theory. In A. Bryant & K. Charmaz (Eds.), The SAGE Handbook of Grounded Theory: Paperback Edition (pp. 239-244): SAGE Publications.

Moulton, C. D., Pickup, J. C., & Ismail, K. (2015). The link between depression and diabetes: the search for shared mechanisms. The Lancet Diabetes & Endocrinology, 3(6), 461-471. doi: 10.1016/S2213-8587(15)00134-5

Nagrebetsky, A., Brettell, R., Roberts, N., & Farmer, A. (2014). Smoking cessation in adults with diabetes: a systematic review and meta-analysis of data from randomised controlled trials. BMJ Open, 4(3). doi: 10.1136/bmjopen-2013-004107

Nam, S., Chesla, C., Stotts, N. A., Kroon, L., & Janson, S. L. (2011). Barriers to diabetes management: patient and provider factors. Diabetes Research and Clinical Practice, 93(1), 1-9. doi: 10.1016/j.diabres.2011.02.002

Nathan, D. M., Buse, J. B., Davidson, M. B., Ferrannini, E., Holman, R. R., Sherwin, R., & Zinman, B. (2009). Medical management of hyperglycemia in type 2 diabetes: a consensus algorithm for the initiation and adjustment of therapy: a consensus statement of the American Diabetes Association and the European Association for the Study of Diabetes. Diabetes Care, 32(1), 193-203. doi: 10.2337/dc08-9025

National Center for Education Statistics. (2014). National Assessment of Adult Literacy (NAAL). Retrieved September 12, 2014, from http://nces.ed.gov/naal/fr_definition.asp

National Committee for Quality Assurance. (2013). HEDIS & Performance Measurement. HEDIS & Quality Measurement. Retrieved May 24, 2013, from http://www.ncqa.org/HEDISQualityMeasurement.aspx

National Guideline Clearinghouse. (2013). Management of type 2 diabetes mellitus. Retrieved April 3, 2013, from http://www.guideline.gov/content.aspx?id=38618

National Institute of Mental Health. (2014). Health & Education. from http://www.nimh.nih.gov/health/index.shtml

Nemeth, L. S., Ornstein, S. M., Jenkins, R. G., Wessell, A. M., & Nietert, P. J. (2012). Implementing and evaluating electronic standing orders in primary care practice:

275

A PPRNet Study. The Journal of the American Board of Family Medicine, 25(5), 594-604. doi: 10.3122/jabfm.2012.05.110214

Newlin, K., Melkus, G. D., Tappen, R., Chyun, D., & Koenig, H. G. (2008). Relationships of religion and spirituality to glycemic control in Black women with type 2 diabetes. Nursing Research, 57(5), 331-339 doi: 10.1097/01.NNR.0000313497.10154.66

Norris, S. L., Lau, J., Smith, S. J., Schmid, C. H., & Engelgau, M. M. (2002). Self-management education for adults with type 2 diabetes: A meta-analysis of the effect on glycemic control. Diabetes Care, 25(7), 1159-1171. doi: 10.2337/diacare.25.7.1159

North Texas Nurse Practitioners. (2015). About Us: General Information. Retrieved September 1, 2014, from https://ntnp.enpnetwork.com/

O'Connor, P. J., Bodkin, N. L., Fradkin, J., Glasgow, R. E., Greenfield, S., Gregg, E., . . . Wysham, C. H. (2011). Diabetes performance measures: current status and future directions. Diabetes Care, 34(7), 1651-1659. doi: 10.2337/dc11-0735

O'Reilly, K. (2012). Ethnographic Methods (2nd ed.): Taylor & Francis. Offredy, M., Kendall, S., & Goodman, C. (2008). The use of cognitive continuum theory

and patient scenarios to explore nurse prescribers' pharmacological knowledge and decision-making. International Journal of Nursing Studies, 45(6), 855-868. doi: 10.1016/j.ijnurstu.2007.01.014

Ogden, J., & Lo, J. (2012). How meaningful are data from Likert scales? An evaluation of how ratings are made and the role of the response shift in the socially disadvantaged. Journal of Health Psychology, 17(3), 350-361. doi: 10.1177/1359105311417192

Ogedegbe, G., Mancuso, C. A., Allegrante, J. P., & Charlson, M. E. (2003). Development and evaluation of a medication adherence self-efficacy scale in hypertensive African-American patients. Journal of clinical epidemiology, 56(6), 520-529. doi: 10.1016/s0895-4356(03)00053-2

Ovretveit, J. (2011). Does clinical coordination improve quality and save money? Full Report. The Health Foundation, 2.

Oxendine, V., Meyer, A., Reid, P. V., Adams, A., & Sabol, V. (2014). Evaluating diabetes outcomes and costs within an ambulatory setting: a strategic approach utilizing a clinical decision support system. Clinical Diabetes, 32(3), 113-120. doi: 10.2337/diaclin.32.3.113

Passmore, C., Doobie, A. E., Parchman, M., & Tysinger, J. (2002). Guidelines for Constructing a Survey. Family Medicine, 34(4), 281-286.

Pearson, J. F., Brownstein, C. A., & Brownstein, J. S. (2011). Potential for electronic health records and online social networking to redefine medical research. Clinical Chemistry, 57(2), 196-204. doi: 10.1373/clinchem.2010.148668

Peyrot, M., Rubin, R. R., & Siminerio, L. M. (2006). Physician and nurse use of psychosocial strategies in diabetes care: results of the cross-national Diabetes

276

Attitudes, Wishes and Needs (DAWN) study. Diabetes Care, 29(6), 1256-1262. doi: 10.2337/dc05-2444

Philis-Tsimikas, A., Gilmer, T. P., Schultz, J., Walker, C., Fortmann, A. L., & Gallo, L. C. (2012). Community-created programs: Can they be the basis of innovative transformations in our health care practice? Implications from 15 years of testing, translating, and implementing community-based, culturally tailored diabetes management programs. Clinical Diabetes, 30(4), 156-163. doi: 10.2337/diaclin.30.4.156

Piette, J. D., & Kerr, E. A. (2006). The impact of comorbid chronic conditions on diabetes care. Diabetes Care, 29(3), 725-731. doi: 10.2337/diacare.29.03.06.dc05-2078

Piette, J. D., Schillinger, D., Potter, M. B., & Heisler, M. (2003). Dimensions of patient-provider communication and diabetes self-care in an ethnically diverse population. Journal of General Internal Medicine, 18(8), 624-633. doi: 10.1046/j.1525-1497.2003.31968.x

Plomin, R., & Asbury, K. (2005). Nature and nurture: genetic and environmental influences on behavior. The ANNALS of the American Academy of Political and Social Science, 600(1), 86-98. doi: 10.1177/0002716205277184

Pogach, L. M., Brietzke, S. A., Cowan, J., Claude L., Conlin, P., Walder, D. J., & Sawin, C. T. (2004). Development of evidence-based clinical practice guidelines for diabetes: the Department of Veterans Affairs/Department of Defense guidelines initiative. Diabetes Care, 27(suppl 2), b82-b89. doi: 10.2337/diacare.27.suppl_2.B82

Polonsky, W. H., Fisher, L., Earles, J., Dudl, R. J., Lees, J., Mullan, J., & Jackson, R. A. (2005). Assessing psychosocial distress in diabetes: development of the diabetes distress scale. Diabetes Care, 28(3), 626-631. doi: 10.2337/diacare.28.3.626

Polzer, R. L., & Miles-Shandor, M. (2005). Spirituality and self-management of diabetes in African Americans. Journal of Holistic Nursing, 23(2), 230-250. doi: 10.1177/0898010105276179

Polzer, R. L., & Miles-Shandor, M. (2007). Spirituality in African Americans with diabetes: self-management through a relationship with God. Qualitative Health Research, 17(2), 176-188. doi: 10.1177/1049732306297750

Powledge, T. M. (2011). Behavioral epigenetics: how nurture shapes nature. BioScience, 61(8), 588-592. doi: 10.1525/bio.2011.61.8.4

Quinn, M. T., Cook, S., Nash, K., & Chin, M. H. (2001). Addressing religion and spirituality in African Americans with diabetes. The Diabetes Educator, 27(5), 643-655. doi: 10.1177/014572170102700505

Radwin, L. E., & Alster, K. (2002). Individualized nursing care: an empirically generated definition. International Nursing Review, 49(1), 54-63. doi: 10.1046/j.1466-7657.2002.00101.x

277

Rasch, R. F. R., & Cogdill, K. W. (1999). Nurse practitioners' information needs and information seeking: implications for practice and education. Holist Nurs Pract, 13(4), 90-97.

Ravitch, S. M., & Riggan, M. (2011). Reason & Rigor: How Conceptual Frameworks Guide Research: SAGE Publications.

Resnick, H. E., Foster, G. L., Bardsley, J., & Ratner, R. E. (2006). Achievement of American Diabetes Association clinical practice recommendations among U.S. adults with diabetes, 1999–2002: the National Health and Nutrition Examination Survey. Diabetes Care, 29(3), 531-537. doi: 10.2337/diacare.29.03.06.dc05-1254

Riddle, M. C., & Karl, D. M. (2012). Individualizing targets and tactics for high-risk patients with type 2 diabetes: practical lessons from ACCORD and other cardiovascular trials. Diabetes Care, 35(10), 2100-2107. doi: 10.2337/dc12-0650

Rintala, T.-M., Jaatinen, P., Paavilainen, E., & Åstedt-Kurki, P. (2013). Interrelation between adult persons with diabetes and their family: a systematic review of the literature. Journal of Family Nursing, 19(1), 3-28. doi: 10.1177/1074840712471899

Ritholz, M. D., Beverly, E. A., Brooks, K. M., Abrahamson, M. J., & Weinger, K. (2014). Barriers and facilitators to self-care communication during medical appointments in the United States for adults with type 2 diabetes. Chronic Illness, 10(4), 303-313. doi: 10.1177/1742395314525647

Robert Wood Johnson Foundation. (2011). Health care's blind side: the overlooked connection between social needs and good health. Retrieved from: http://www.rwjf.org/content/dam/farm/reports/surveys_and_polls/2011/rwjf71795

Robinson, J. W., & Roter, D. L. (1999a). Counseling by primary care physicians of patients who disclose psychosocial problems. J Fam Pract, 48(9), 698-705.

Robinson, J. W., & Roter, D. L. (1999b). Psychosocial problem disclosure by primary care patients. Social Science & Medicine, 48(10), 1353-1362. doi: 10.1016/S0277-9536(98)00439-0

Rodbard, H. W., Jellinger, P. S., Davidson, J. A., Einhorn, D., Garber, A. J., Grunberger, G., . . . Schwartz, S. S. (2009). Statement by an American Association of Clinical Endocrinologists/American College of Endocrinology Consensus Panel on type 2 diabetes mellitus: an algorithm for glycemic control. Endocrine Practice, 15(6), 540-559. doi: 10.4158/EP.15.6.540

Rogowski, J., Lillard, L. A., & Kington, R. (1997). The financial burden of prescription drug use among elderly persons. The Gerontologist, 37(4), 475-482. doi: 10.1093/geront/37.4.475

Russell, L. B., Suh, D.-C., & Safford, M. A. (2005). Time requirements for diabetes self-management: too much for many? The Journal of Family Practice, 54(1), 52-56.

Rutten, G. E. H. M. (2005). Diabetes patient education: time for a new era. Diabetic Medicine, 22, 671-673.

278

Rycroft-Malone, J., Seers, K., Titchen, A., Harvey, G., Kitson, A., & McCormack, B. (2004). What counts as evidence in evidence-based practice? Journal of Advanced Nursing, 47(1), 81-90. doi: 10.1111/j.1365-2648.2004.03068.x

Saaddine, J. B., Engelgau, M. M., Beckles, G. L., Gregg, E. W., Thompson, T. J., & Narayan, K. M. V. (2002). A diabetes report card for the United States: quality of care in the 1990s. Annals of Internal Medicine, 136(8), 565-574. doi: 10.7326/0003-4819-136-8-200204160-00005

Safford, M. M., Russell, L., Suh, D.-C., Roman, S., & Pogach, L. (2005). How much time do patients with diabetes spend on self-care? The Journal of the American Board of Family Practice, 18(4), 262-270. doi: 10.3122/jabfm.18.4.262

Saldaña, J. (2009). The Coding Manual for Qualitative Researchers: SAGE Publications. Sampson, R. J., Morenoff, J. D., & Raudenbush, S. (2005). Social anatomy of racial and

ethnic disparities in violence. American Journal of Public Health, 95(2), 224-232. doi: 10.2105/AJPH.2004.037705

Scheder, J. C. (1988). A sickly-sweet harvest: farmworker diabetes and social equality. Medical Anthropology Quarterly, 2(3), 251-277. doi: 10.1525/maq.1988.2.3.02a00050

Schensul, J. J., & LeCompte, M. D. (2012). Essential Ethnographic Methods: A Mixed Methods Approach (2nd ed. Vol. 3): Rowman Altamira.

Schillinger, D., Grumbach, K., Piette, J., Wang, F., Osmond, D., Daher, C., . . . Bindman, A. B. (2002). Association of health literacy with diabetes outcomes. The Journal of the American Medical Association, 288(4), 475-482. doi: 10.1001/jama.288.4.475

Schillinger, D., Grumbach, K., Piette, J., Wang, F., Osmond, D., Daher, C., . . . Bindman, A. B. (2002). Association of health literacy with diabetes outcomes. The Journal of the American Medical Association, 288(4), 475-482. doi: 10.1001/jama.288.4.475

Schmittdiel, J. A., Uratsu, C. S., Karter, A. J., Heisler, M., Subramanian, U., Mangione, C. M., & Selby, J. V. (2008). Why don't diabetes patients achieve recommended risk factor targets? Poor adherence versus lack of treatment intensification. Journal of General Internal Medicine, 23(5), 588-594. doi: 10.1007/s11606-008-0554-8

Selby, J. V. (2010). Health systems, patients factors, and quality of care for diabetes: A synthesis of findings from the TRIAD study. Diabetes Care, 33(4), 940-947. doi: 10.2337/dc09-1802

Senteio, C., & Veinot, T. C. E. (2014). Trying to make things right: adherence work in high-poverty, African American neighborhoods. Qualitative Health Research, 24(12), 1745-1756. doi: 10.1177/1049732314549027

Shachak, A., & Reis, S. (2009). The impact of electronic medical records on patient–doctor communication during consultation: a narrative literature review. Journal

279

of Evaluation in Clinical Practice, 15(4), 641-649. doi: 10.1111/j.1365-2753.2008.01065.x

Sharfi, R. G., Azad, B. L., Avat, F., & Siamak, M. (2013). Importance of social support in diabetes care. Journal of Education and Health Promotion, 2(62).

Sharma, V. K., Lepping, P., Cummins, A. G. P., Copeland, J. R. M., Parhee, R., & Mottram, P. (2004). The Global Mental Health Assessment Tool - Primary Care Version (GMHAT/PC). development, reliability and validity. World Psychiatry, 3(2), 115-119.

Shenassa, E. D., Stubbendick, A., & Brown, M. J. (2004). Social disparities in housing and related pediatric injury: a multilevel study. American Journal of Public Health, 94(4), 633-639. doi: 10.2105/AJPH.94.4.633

Sherman, K., Eaves, E., Ritenbaugh, C., Hsu, C., Cherkin, D., & Turner, J. (2014). Cognitive interviews guide design of a new CAM patient expectations questionnaire. BMC Complementary and Alternative Medicine, 14(1), 39. doi: 10.1186/1472-6882-14-39

Shield, R. R., Goldman, R. E., Anthony, D. A., Wang, N., Doyle, R. J., & Borkan, J. (2010). Gradual electronic health record Implementation: new insights on physician and patient adaptation. The Annals of Family Medicine, 8(4), 316-326. doi: 10.1370/afm.1136

Shier, G., Ginsburg, M., Howell, J., Volland, P., & Golden, R. (2013). Strong social support services, such as transportation and help for caregivers, can lead to lower health care use and costs. Health affairs, 32(3), 544-551. doi: 10.1377/hlthaff.2012.0170

Shrivastava, S., Shrivastava, P., & Ramasamy, J. (2013). Role of self-care in management of diabetes mellitus. Journal of Diabetes & Metabolic Disorders, 12(1), 14. doi: 10.1186/2251-6581-12-14

Siminerio, L. M., Funnell, M. M., Peyrot, M., & Rubin, R. R. (2007). U.S. nurses' perceptions of their role in diabetes care: results of the cross-national diabetes attitudes wishes and needs (DAWN) study. The Diabetes Educator, 33(1), 152-162. doi: 10.1177/0145721706298194

Siminerio, L. M., Wagner, E. H., Gabbay, R., & Zgibor, J. (2009). Implementing the chronic care model: A statewide focus on improving diabetes care for Pennsylvania. Clinical Diabetes, 27(4), 153-159. doi: 10.2337/diaclin.27.4.153

Singh-Manoux, A. (2003). Psychosocial factors and public health. Journal of Epidemiology and Community Health, 57(8), 553-556. doi: 10.1136/jech.57.8.553

Skovlund, S. E., & Peyrot, M. (2005). The Diabetes Attitudes, Wishes, and Needs (DAWN) Program: A new approach to improving outcomes of diabetes care. Diabetes Spectrum, 18(3), 136-142. doi: 10.2337/diaspect.18.3.136

Snow, V., Aronson, M. D., Hornbake, E. R., Mottur-Pilson, C., & Weiss, K. B. (2004). Lipid control in the management of type 2 diabetes mellitus: A clinical practice

280

guideline from the American College of Physicians. Annals of Internal Medicine, 140(8), 644-649. doi: 10.7326/0003-4819-140-8-200404200-00012

Song, Y., Song, H.-J., Han, H.-R., Park, S.-Y., Nam, S., & Kim, M. T. (2012). Unmet needs for social support and effects on diabetes self-care activities in Korean Americans with type 2 diabetes. The Diabetes Educator, 38(1), 77-85. doi: 10.1177/0145721711432456

Soto, S. C., Louie, S. Y., Cherrington, A. L., Parada, H., Horton, L. A., & Ayala, G. X. (2015). An ecological perspective on diabetes self-care support, self-management behaviors, and hemoglobin A1C among Latinos. Diabetes Educ, 41(2), 214-223. doi: 10.1177/0145721715569078

Spector, P. E. (2004). Likert Scale. In M. E. Lewis-Beck, A. Bryman, & T. F. Liao (Eds.), The SAGE Encyclopedia of Social Science Reseach Methods.

Spencer, M. S., Rosland, A.-M., Kieffer, E. C., Sinco, B. R., Valerio, M., Palmisano, G., . . . Heisler, M. (2011). Effectiveness of a community health worker intervention among African American and Latino adults with type 2 diabetes: A randomized controlled trial. American Journal of Public Health, 101(12), 2253-2260. doi: 10.2105/ajph.2010.300106

Spoont, M., Greer, N., Su, J., Fitzgerald, P., Rutks, I. B., & Wilt, T. J. (2011). Rural vs. urban ambulatory health care: a systematic review. Retrieved from: http://www.ncbi.nlm.nih.gov/books/NBK56144/#!po=16.6667

Squires, J. E., Hayduk, L., Hutchinson, A. M., Cranley, L. A., Gierl, M., Cummings, G. G., . . . Estabrooks, C. A. (2013). A protocol for advanced psychometric assessment of surveys. Nurs Res Pract, 2013, 156782. doi: 10.1155/2013/156782

Strayer, S. M., Shaughnessy, A. F., Yew, K. S., Stephens, M. B., & Slawson, D. C. (2010). Updating clinical knowledge: An evaluation of current information alerting services. International journal of medical informatics, 79(12), 824-831. doi: 10.1016/j.ijmedinf.2010.08.004

Syed, S. T., Gerber, B. S., & Sharp, L. K. (2013). Traveling towards disease: transportation barriers to health care access. Journal of Community Health, 38(5), 976-993. doi: 10.1007/s10900-013-9681-1

Tang, T. S., Ayala, G. X., Cherrington, A., & Rana, G. (2011). A review of volunteer-based peer support interventions in diabetes. Diabetes Spectrum, 24(2), 85-98. doi: 10.2337/diaspect.24.2.85

Tang, T. S., Brown, M. B., Funnell, M. M., & Anderson, R. M. (2008). Social support, quality of life, and self-care behaviors among African Americans with type 2 diabetes. The Diabetes Educator, 34(2), 266-276. doi: 10.1177/0145721708315680

Tapp, H., Phillips, S. E., Waxman, D., Alexander, M., Brown, R., & Hall, M. (2012). Multidisciplinary team approach to improved chronic care management for diabetic patients in an urban safety net ambulatory care clinic. The Journal of the American Board of Family Medicine, 25(2), 245-246. doi: 10.3122/jabfm.2012.02.110243

281

Tastle, W. J., & Wierman, M. J. (2006). An information theoretic measure for the evaluation of ordinal scale data. Behavior Research Methods, 38(3), 487-494. doi: 10.3758/BF03192803

Tavakol, M., & Dennick, R. (2011). Making sense of Cronbach's alpha. International Journal of Medical Education, 2, 53-55. doi: 10.5116/ijme.4dfb.8dfd

Taylor, R. S. (1982). Value-added processes in the information life cycle. Journal of the American Society for Information Science, 33(5), 341-346. doi: 10.1002/asi.4630330517

Taylor, R. S. (1991). Information Use Environments. In B. Dervin (Ed.), Progress in Communication Sciences (Vol. 10, pp. 217-255). Norwood, NJ: Ablex.

Thoits, P. A. (2010). Stress and health: major findings and policy implications. Journal of Health and Social Behavior, 51(1 suppl), S41-S53. doi: 10.1177/0022146510383499

Thompson, C., Cullum, N., McCaughan, D., Sheldon, T., & Raynor, P. (2004). Nurses, information use, and clinical decision making—the real world potential for evidence-based decisions in nursing. Evidence-Based Nursing, 7, 68-72. doi: 10.1136/ebn.7.3.68

Thorpe, C., Ryan, B., McLean, S. L., Burt, A., Stewart, M., Brown, J. B., . . . Harris, S. (2009). How to obtain excellent response rates when surveying physicians. Family Practice, 26(1), 65-68. doi: 10.1093/fampra/cmn097

Tierney, W. M., Overhage, J. M., & McDonald, C. J. (1996). Computerizing guidelines: factors for success. American Medical Informatics Association.

Todd, W. E. (1998). Clinical practice guidelines. Auckland ; Philadelphia: ADIS International Limited.

Tulloch-Reid, M., & Williams, D. E. (2003). Quality of diabetes care in the United States between 1988 and 1995. Clinical Diabetes, 21(1), 43-45. doi: 10.2337/diaclin.21.1.43

Two Feathers, J., Kieffer, E. C., Palmisano, G., Anderson, M., Sinco, B., Janz, N., . . . James, S. A. (2005). Racial and ethnic approaches to community health (REACH) Detroit partnership: improving diabetes-related outcomes among African American and Latino adults. American Journal of Public Health, 95(9), 1552-. doi: 10.2105/AJPH.2005.066134

U.S. Census Bureau. (2014). Health Insurance - CPS Health Insurance Definitions. Health Insurance. Last Revised: September 12, 2012. Retrieved April 5, 2014, from https://www.census.gov/hhes/www/hlthins/methodology/definitions/cps.html

Uchino, B. N., Cacioppo, J. T., & Kiecolt-Glaser, J. K. (1996). The relationship between social support and physiological processes: a review with emphasis on underlying mechanisms and implications for health. Psychological Bulletin, 119(3), 488-531.

United States Census Bureau. (2014). Retrieved September 6, 2014, from https://www.census.gov/hhes/www/hlthins/methodology/definitions/cps.html

282

van Dam, H. A., van der Horst, F., van den Borne, B., Ryckman, R., & Crebolder, H. (2003). Provider–patient interaction in diabetes care: effects on patient self-care and outcomes. Patient Education and Counseling, 51(1), 17-28. doi: 10.1016/s0738-3991(02)00122-2

VanGeest, J. B., Johnson, T. P., & Welch, V. L. (2007). Methodologies for improving response rates in surveys of physicians: a systematic review. Evaluation & the Health Professions, 30(4), 303-321. doi: 10.1177/0163278707307899

Veinot, T. C. E., Zheng, K., Lowery, J. C., Souden, M., & Keith, R. (2010). Using Electronic Health Record Systems in Diabetes Care: Emerging Practices. Paper presented at the Health Informatics Symposium (IHI'10), Arlington, Virginia, USA.

Ventres, W., Kooienga, S., Vuckovic, N., Marlin, R., Nygren, P., & Stewart, V. (2006). Physicians, patients, and the electronic health record: an ethnographic analysis. The Annals of Family Medicine, 4(2), 124-131. doi: 10.1370/afm.425

Versnel, N., Welschen, L., Baan, C., Nijpels, G., & Schellevis, F. (2011). The effectiveness of case management for comorbid diabetes type 2 patients; the CasCo study. Design of a randomized controlled trial. BMC Family Practice, 12(1), 68. doi: 10.1186/1471-2296-12-68

Wagner, D. V., Stoeckel, M., Tudor, M. E., & Harris, M. A. (2015). Treating the most vulnerable and costly in diabetes. Curr Diab Rep, 15(6), 606. doi: 10.1007/s11892-015-0606-5

Wagner, J., & Abbott, G. (2007). Depression and depression care in diabetes: relationship to perceived discrimination in African Americans. Diabetes Care, 30(2), 364-366. doi: 10.2337/dc06-1756

Walker, R. E., Keane, C. R., & Burke, J. G. (2010). Disparities and access to healthy food in the United States: A review of food deserts literature. Health & Place, 16(5), 876-884. doi: 10.1016/j.healthplace.2010.04.013

Walsh, S. H. (2004). The clinician’s perspective on electronic health records and how they can affect patient care. British Medical Journal, 328, 1184-1187.

Walter, L. C., Davidowitz, N. P., Heineken, P. A., & Covinsky, K. E. (2004). Pitfalls of converting practice guidelines into quality measures: Lessons learned from a VA performance measure. Journal of the American Medical Association, 291(20), 2466-2470. doi: 10.1001/jama.291.20.2466

Walton, J. W., Snead, C. A., Collinsworth, A. W., & Schmidt, K. L. (2012). Reducing diabetes disparities through the implementation of a community health worker–led diabetes self-management education program. Family & Community Health, 35(2), 161-171. doi: 10.1097/FCH.0b013e31824651d3

Watkins, Y. J., Quinn, L. T., Ruggiero, L., Quinn, M. T., & Choi, Y.-K. (2013). Spiritual and religious beliefs and practices and social support’s relationship to diabetes self-care activities in African Americans. The Diabetes Educator, 39(2), 231-239. doi: 10.1177/0145721713475843

283

Weiner, S. J. (2004). Contextualizing medical decisions to individualize care. Journal of General Internal Medicine, 19(3), 281-285. doi: 10.1111/j.1525-1497.2004.30261.x

Weiner, S. J., Kelly, B., Ashley, N., Binns-Calvey, A., Sharma, G., Schwartz, A., & Weaver, F. M. (2014). Content coding for contextualization of care: evaluating physician performance at patient-centered decision making. Medical Decision Making, 34(1), 97-106. doi: 10.1177/0272989x13493146

Weiner, S. J., Schwartz, A., Weaver, F., Goldberg, J., Yudkowsky, R., Sharma, G., . . . Abrams, R. I. (2010). Contextual errors and failures in individualizing patient care: a multicenter study. Annals of Internal Medicine, 153(2), 69-75. doi: 10.7326/0003-4819-153-2-201007200-00002

Wideman, T. H., & Sullivan, M. J. L. (2012). Development of a cumulative psychosocial factor index for problematic recovery following work-related musculoskeletal injuries. Physical Therapy, 92(1), 58-68. doi: 10.2522/ptj.20110071

Wiebe, E. R., Kaczorowski, J., & MacKay, J. (2012). Why are response rates in clinician surveys declining? Canadian Family Physician, 58(4), e225-e228.

Wild, D., von Maltzahn, R., Brohan, E., Christensen, T., Clauson, P., & Gonder-Frederick, L. (2007). A critical review of the literature on fear of hypoglycemia in diabetes: Implications for diabetes management and patient education. Patient Education and Counseling, 68(1), 10-15. doi: 10.1016/j.pec.2007.05.003

Willens, D., Cripps, R., Wilson, A., Wolff, K., & Rothman, R. (2011). Interdisciplinary team care for diabetic patients by primary care physicians, advanced practice nurses, and clinical pharmacists. Clinical Diabetes, 29(2), 60-68. doi: 10.2337/diaclin.29.2.60

Willis, G. B. (2005). Cognitive Interviewing: A Tool for Improving Questionnaire Design: SAGE Publications.

Wilson, P. M., Petticrew, M., Calnan, M., & Nazareth, I. (2010). Effects of a financial incentive on health researchers’ response to an online survey: a randomized controlled trial. Journal of Medical Internet Research, 12(2), e13. doi: 10.2196/jmir.1251

Wilson, W., Ary, D. V., Biglan, A., Glasgow, R. E., Toobert, D. J., & Campbell, D. R. (1986). Psychosocial predictors of self-care behaviors (compliance) and glycemic control in non-insulin-dependent diabetes mellitus. Diabetes Care, 9(6), 614-622. doi: 10.2337/diacare.9.6.614

Wright, A., McCoy, A. B., Henkin, S., Kale, A., & Sittig, D. F. (2013). Use of a support vector machine for categorizing free-text notes: assessment of accuracy across two institutions. Journal of the American Medical Informatics Association, 20(5), 887-890. doi: 10.1136/amiajnl-2012-001576

Zhou, X., Ackerman, M. S., & Zheng, K. (2009). I just don’t know why it’s gone: maintaining informal information use in inpatient care. Paper presented at the Computer Human Interaction Conference - CHI 2009, Boston, MA.

284

http://www.lib.umich.edu/articles/details/FETCH-LOGICAL-a433-d3e41bb551e324a0b9bbf59d12c8d7f2c36bbb595c03447f539df44b4003fc6a3

Zhou, X., Ackerman, M. S., & Zheng, K. (2010). Doctors and psychosocial information: records and reuse in inpatient care. Paper presented at the Proceedings of the SIGCHI Conference on Human Factors in Computing Systems, Atlanta, Georgia, USA. http://dl.acm.org/citation.cfm?id=1753592